New Data Center Standard

mstansberry writes to tell us that the Telecommunications Industry Association (the people who brought you the CAT standards for unshielded twisted pair cabling) recently published a 148 page document meant to standardize the design considerations for every single aspect of a data center. The standard covers everything from site selection to rack mounting methods.

Comments

twisting

[seanadams.com]

Speaking of CAT standards, has anyone else had a good look at the differences between CAT3, CAT5, etc?

CAT5 just seems to be twisted a little tighter, but CAT6 actually modifies the twist gradually, in a cycle that repeats every few feet, with each pair 90 degrees "out of phase" from the next. Plus theres (sometimes) a plastic "spine" in there to maintain spacing and/or bend radius. It's not obvious to me how varying the twists-per-foot along the cable should help - anyone know?

Re:twisting

[Mr. Flibble]

because of the extra twisting there is less crosstalk. The less crosstalk, the more information you can pump down the wire, thus the difference between Cat 3,5 and 6.

(FYI, crosstalk is interference from a parralell channel in the wire)

Re:twisting

[rusty0101]

Not sure entirely myself, however as a thought, having a constant twist (different from one pair to the next, but the same for the length of the cable) could set up a situation where two cables are in phase at several points alog the cable run, and some signal transfer may hapen.

Varying the twist rate along the run of a pair, as well as doing what you can to keep it out of phase with other pairs by braiding, or other means would make it possible to set up a longer cable run without viable phase transfer points that could cause signal bleed between pairs.

However that's just conjecture on my part. I am sure someone will come along who can give us the math to show that my conjecture is entirely wrong.

-Rusty

Re:twisting

[william_w_bush]

reduces the ability for coupling between strands at a regular point. essentially spreading the crosstalk from a different pair between different coupling points. instead of a pair being completely coupled or not the coupling is spread such that 1 whole interference cycle is generated, cancelling out all crosstalk, or at least more than the constant twist.

tightly twisted pairs take longer to go through than loosely twisted pairs, and so instead of 2 parallel antennae inducting into each other you have two antellae parallel at different points in such a way that the inducted wave at one portion of the cable destructively interferes with the inducted wave at a different part of the cable.

Re:twisting

[SoloFlyer2]

Varying the twist rate only helps when u have several cables together ( Think 4 or 5 cables in a conduit ). Basically it just reduces the chance that 2 cables next to each other are going to have exactly the same twist rate.

In other words it reduces cross-talk between cables :)

Re:twisting

[boisepunk]

You sir are to be commended just for the fact that you have admitted some of your content is conjecture. I'm going to add you to my friends list because of this. You have earned my respect.

Re:twisting

[mhearne]

Visibly, 10t, cat5, cat5e and cat6 don't seem much different do they?

I think that it has everything to do with the dielectric of the insulation, rather than some magic twist in the wire.

Michael

Re:twisting

[william_w_bush]

oh fyi, this doesn't help at lower speeds, but common lvds serdes nowadays hit upwards of 10Gbps, so this kind of thing makes a big difference at that point because the wavelength is small enough that the difference becomes possible to vary in the cabling.

basically for 1gbps its just starting, past 5gbps (10g copper spec calls for this though a working group is trying to run 10g on cat5 (boggle)) you need this kind of thing, and it happens to cut down on rf interference and reception, another handy thing, considering how hard it is to get some of these things to pass fcc (been there, done that).

Re:twisting

[DaEMoN128]

I am BICC certified. There is a difference between CAT3 and CAT 5. The twists are much more pronounced on CAT 5. It may not look like it, but there is a big differnece. There is also different amounts of twists per pair. CAT 6 is something I havent had much hands on with (we use mostly fiber for that stuff). CAT 3 has much lower frequency response per pair than CAT 5. A good cable tester can actually verify that for me if you can get your hands on one (see network analyser). Just from looking online, CAT 6 has a minimum 250MHz bandwith while CAT 5 has a minimum 100 MHz bandwith per pair. http://www.lanshack.com/cat5e-tutorial.asp check here. It says CAT 5e is the same as CAT 6 but CAT 6 is manufactured to a higher standard. I guess that the fab tolerances are tighter for Cat 6.

Re:twisting

[william_w_bush]

also yes and no, the added twisting makes the crosstalk elimination better at different freqencies. at 1ghz a single turn is enough to cause the antenna effect, while at 10mhz (original spec) you need like half a meter or something.

Re:twisting

[Fnord]

Actually it has everything to do with the twist in the wire. If you twist a +/- wire pair together, the signals they put out are cancelled some. The tighter the twist the higher frequency is cancelled. I'm not sure the exact reason, but varying the twist along the length of cable could be to remove any resonance.

Re:twisting

[jacksonj04]

Visibly, mains cable and antenna cable don't look that different. Still, I know which one I'd rather cut through by accident.

As it happens, the varying twist rate helps prevent cross-talk, reduces the effect of the cable acting as an antenna, and generally looks fairly cool. The dialectric of the insulation doesn't really feature, unless you're talking about shielded stuff.

Re:twisting

[lithium100]

I'm no expert but fundamentally it comes down to Maxwell's equations. A current is induced by a magnetic field in a wire by a changing magnetic field - ie; an ac signal of some frequency. Similarly, a changing electrical current in a wire will create a changing magnetic field and hence "crosstalk".

There is one caveat though and that is only the orthogonal components (to the wire) of the magnetic field induce a current and so by twisting the wires you minimize the orthogonal components.

At least that what I remember from fields studies back at uni - has been a while!

Re:twisting

[lithium100]

Take a look at a CAT5 cable. There is practically no insulation at all so the dielectric permitivity would be hardly any better than air. Thats WHY they started twisting it so that heavy insulation wasn't needed as it is in coax.

Re:twisting

[baadger]

Additionally, twisting 2 wires together means interference from outside sources (like 50 or 60Hz mains hum?) is made as uniform as possible across them. If you put the same signal onto these 2 wires half a cycle out of phase then you can put the signal at the other end through some simple electronics to produce the difference between them and reject the noise. I would have thought that this is an equally important aspect of the twisting.

Does common mode rejection still play a part in modern ethernet or other networks? (from some CAT5 pinouts it looks to me that it may)

Re:twisting

[budgenator]

the diameter of the insulater arround the center conductor of a coaxial cable has a lot to do with the capacitance of the cable, such as the center conductor is one plate of the capacitor and the shield is the other. The capacitance is needed to cancel out the inductance of the cable, the product of the cable's capacitance, inductance and resistivity gives a cable a characteristic impeadence such as a 75 ohm cable for your TV.

When all of the impeadances match, the power or energy (can't remember which, and they are different technicaly) is maximised.
One of the radars we used to have used a coax cable to the feed horn, the center conductor was about 1/8 copper pipe surrounded by about a 1 1/2 inches of insulation.

Same is true for twisted pairs, the thickness of the insulation effects the capacitance of the pair and the adjacent pairs, smaller diameter insulation gets you more cross-talk at higher frequencies.

Re:twisting

[lithium100]

Its both power and energy. When the impedences match both the voltage across each impedence and the current though it is maximised thus maximising power. Power is simply the rate of change of energy in time. eg; a watt is a joule per second :)

Re:twisting

[FerretFrottage]

Now you're taking about balanced signals which as you mentioned take "some simpe electronics" which takes "some more money". I would be interesting to see what difference there would be in error rates between balanced and unbalanced commuinications over CAT5

Re:twisting

[MECC]

More twists means less near end crosstalk. Cat 6 also specifies minium bend radiuses. Something a girl I once made out with paid attention to.

A good article on cat 6 can be found at The Data Center Journal It points out things like why you don't want to use anything that would clamp down to tightly on the Cat 6 runs, like nylon tie-wraps. Use velcro instead. That said, I've run 1Gbps through home-made Cat5 with no errors, albiet for short distances (less than 50 feet).

Still, $250 just to read a standard. Not good.

Re:twisting

[Sgt_Jake]

Cat 6 is supposed to do gig speeds (1000), cat 5 won't be reliable over 100. or so they say... All of it makes great speaker wire - MUCH better than the cheap but impressive looking fatties at circuit city.

Re:twisting

[Mecha[drone]]

CAT 6 isn't is real standard, the terms Cat 5e and Cat 6 are for GigE purposes interchangable, and are marketing designations.

Cat 3 is almost impossible to use for 100Mb over a distance, but 10Mb is ok if its really Cat3. Cat 3 is typically used for phone Service.

Re:twisting

[ajs318]

When you pull a bell-rope, it stretches slightly. Then the stretched bit shrinks back to how it was and a higher-up bit of the rope is stretched. The stretched bit works its way up the rope to the wheel, which only moves when the last bit of rope snaps back. All this happens far too fast for you to see, but it does happen. {You might be able to see it in a Slinky spring loosely stretched out, especially if you have a camcorder that can do slow motion playback}. Due to the physical properties of the rope and the mechanism at the bell end, there is a small but finite delay between pulling the rope and the bell ringing.

The same thing happens with electrical signals. They do not travel along wires instantaneously. There is an entire subdiscipline of electrical engineering dedicated to the study of transmission lines, but here are some of the points relevant to this discussion.

Every conductor carrying a current radiates a magnetic field. If you could have two wires absolutely coincident and carrying the same current in opposite directions, then the fields they radiate will cancel out exactly and thus not interfere with anything. Obviously you can't get them absolutely coincident, because real matter takes up space. But it turns out that intimate proximity is good enough for real world applications. So, we twist the wires to keep them together ..... this way we don't have to have a huge bulky insulation jacket forcing them together, and can fit more wires into the same space. We can help matters by ensuring that the load on the far end is perfectly resistive: i.e. that every bit of energy put into it actually does work {or just gets turned into heat} rather than getting stored and released.

Now, any wire that is not absolutely straight looks like a coil. Electrons travelling along a coil behave as though they have inertia {due to energy being stored as a magnetic field and then released} and the faster they are moving, the harder it is to persuade them to change direction. If you have a coil with many turns and a socking great lump of steel up the middle, it takes awhile to build up a current from a battery because energy is being stored in magnetising the steel. {Water flow analogy: imagine a turbine in the pipe with a heavy flywheel. It takes an effort to get it up to speed, and it wants to keep turning -- and trying to shove water along -- even after the water stops.} When there are no more molecules to line up with the field, the core is saturated and the coil behaves as a simple resistor until the current changes. Disconnect the battery, and the electrons will actually take some time to come to a halt. That's why you often see a diode -- and maybe a resistor in series with the diode, if it's a really monstrously inductive load like a railway train motor* -- across a relay or motor switched by a transistor. The diode gives the current a path to flow through without having to arc across {a pure current source without a resistor in parallel with itself behaves as an infinite voltage source; most things become good conductors if you apply a high enough voltage}. Air obviously cannot store as much energy in the form of magnetism as steel; but at a high enough frequency, the current will have changed direction before magnetic saturation occurs. Which is why VHF inductances usually have only a few turns on a piddly little ferrite core, and UHF inductances often are literally just bent bits of wire. {Of course, these can even be printed}.

Twisting two wires together actually forms a transformer. It's not a brilliant one, but it will work as one at high frequencies. Each wire induces a current in the other. The induced currents should cancel one another out perfectly. But of course two wires twisted together also form a capacitor. So your twisted pair cable is actually this huge inductive, capacitive, resistive nightmare of a thing.

A cable also has a characteristic impedance. When you put a battery to one end, a pulse current flow

Re:twisting

[Anonymous Coward]

How do you wire it? Do you only use single wires or do you use multiple pairs per speaker?

Re:twisting

[networkBoy]

All ethernet comms is balenced (differential).
There are two ways to remove the common mode signal from a differential pair:
1) an opAmp
2) a balun transformer.

The reason why manufacturers use number 1 is because it is cheaper, tunable, and weighs less.
-nB

Re:twisting

[wings]

Here's a link that discusses the measurements done on CAT cables:

http://whatis.techtarget.com/definition/0,,sid9_gc i1059969,00.html

Re:twisting

[wings]

I should have included this link too. It shows a comparison of the specs for CAT 5 through CAT6:

http://discountcablesusa.com/ethernet-cables.html

Re:twisting

[budgenator]

I wasn't sure, in non-electric physics it was power = M*V and energy was e = M*V^2, it can make quite a difference in what kind of bullet your being shot with, M-14 high power, lower energy vs. M-16 lower power, higher energy; M-16 is less likely to go clean through so you absorb all the energy, ouch.

Re:twisting

[Kevin DeGraaf]

Not exactly.

Cat 3 - 10 Mbps
Cat 5 - 100 Mbps
Cat 5e - 1,000 Mbps (1 Gbps)
Cat 6 - 10,000 (10 Gbps)

Re:twisting

[Sgt_Jake]

riiiight - thanks!

Re:twisting

[mhearne]

Ok. Most of the signal is going to exist on the outside of the wire (the skin effect). Does a stranded wire have more surface area than a solid wire?

Michael

Re:twisting

[mhearne]

I have seen a shielded lan cable, it was very old and unwieldy (IBM used to make some real monster equipment, when clock frequencies were near TV/Radio bands).

I have never seen a shielded cat5 or cat5e. You are the second person that didn't thing much about my dielectric theory, but what else could it be?

If the twist is really the answer, then we must be talking about pico-inductances. It seems to me that inductances that small would only effect the mid-microwave range (20-50 GHz).

Well, this may getting moot. As long as thing works, I guess I'll just be happy with it. Amazing, isn't it, we've gone from striking a pair of rocks together to build a fire, to all that we have today, in only a few thousands of years.

Let's not forget our roots, in case the time comes when we need to start over.

Michael

Re:twisting

[mhearne]

There are different dielectrics. For instance, regular rg-8 cable is about 5/16-inch in diameter, while mini rg-8 is 3/16-inch. But they are electrically equivalent.

Similarly, a large 100 microfarad electrolytic capactitor can be reduced to the size of a pencil eraser by using tantalum or mylar - I have tuned microstrips with mylar tape, which had no electrical connection to anything.

I won't try to get to deep here, because I'm not really qualified. I worked as an assembler on the awacs, not an engineer.

By the way:
Power = Voltage x Current and
Power = Current squared x Resistance

so a power source delivers maximum output to the load when the impedance of load equals the impedance of the source. This is the concept behind using an antenna tuner to match electrical, rather than physical characteristics.

Even a transmitter with a fixed frequency needs to have the antenna matched to the transmitter, there is no *perfect* antenna.

Michael

Not this again ...

[Trusty Penfold]

How boring ... who wants to work somewhere identical to the last place. And identical to your friends' places of work.

How about letting a bit of originality in once in a while?

Oh yeah ... and I'm not bloody paying $250 just to make more work for myself.

---
jon_edwards@spanners4us.com

Re:Not this again ...

[WAG24601G]

I'm not bloody paying $250 just to make more work for myself.

I don't imagine this would be an especially effective means of encouraging compliance either... "we'll tell you what to do for the low, low price of $250!!"

Maybe they should just distribute the literature for free to management-types and then start a rumor of a $250 opt-out fee! (rather like the phone-companies' approach to regional phonebooks)

Re:Not this again ...

[mnemonic_]

Welcome to the real world, where designs that work are more important than "originality." I'm guessing you're one of those people who doesn't like newspapers because of all the consistently black text. "Wouldn't it be nicer if they used different colors?" etc.

Oh, and if you didn't notice, this "standard" is to be used as a recommendation. It's not law. "Letting in originality" isn't an issue when nothing is being forbidden in the first place.

Re:Not this again ...

[Anonymous Coward]

Actually standards help in the purchasing dept.. when you want to rent a couple of racks, you can have a bunch of datacenters place bids on your project, and they'd have to meet certain specs, like amount of cooling, amount of electricity, how many racks, etc.

Re:Not this again ...

[drsmithy]

I don't imagine this would be an especially effective means of encouraging compliance either... "we'll tell you what to do for the low, low price of $250!!"

Out of the costs involved in setting up a data centre, $250 wouldn't even qualify as a rounding error.

Re:Not this again ...

[datafr0g]

How boring ... who wants to work somewhere identical to the last place. And identical to your friends' places of work.

When designing a Data Centre, I really don't think the number one priority is to make it an artistic statement or a fun place for the IT staff to hang out in.

Re:Not this again ...

[Anonymous Crowhead]

When designing a Data Centre, I really don't think the number one priority is to make it an artistic statement or a fun place for the IT staff to hang out in.

Hey, if the 1990s taught us anything, it was that bean bags, nerf guns and video games were the key to a successful business.

Re:Not this again ...

[Jekler]

I can't tell if you're being serious. Assuming you're being serious...

Originality and creativity have certain places in the world. Just because you have guidelines and standards doesn't mean you can't be creative. Programming languages have standards, that doesn't mean programmers can't create original programs. If there were no coding conventions and standards, you'd almost never be able to examine someone elses code. "Wait a second, why are all the integer variables stored as strings? And I think this guy uses + signs as assignment operators..."

Standards are there to prevent entropy, not prevent innovation. Every writer, producer, actor etc. has to write a proposal, film a pilot, and go through all the same stages that everyone else has to. Joss Wheddon, Chris Carter, Jon Stewart... they all had to adhere to the same process that their predecessors did, that didn't stop them from being original and succeeding, it prevented their ideas from being ignored because the people reading the proposal didn't need to interpret it, they already knew the format it was going to be in.

Standards make sure that our ideas are understood universally. In the case of a data center, it ensures that we can all store and retrieve data in a unified way. All the people who need that data don't need to figure out some proprietary system. When you contract services from a standards-compliant center, you don't need to hear the line "That might be what you're used to but... this is OUR way of doing things."

Re:Not this again ...

[Skapare]

The $250 is well above the cost of reproduction, so they are obviously trying to recoup costs of something. Could that be the research done to develop this document? Given some description of its applicability (e.g. it's a PHB checklist) I would tend to think a lot of money was wasted ($250 minus about $30 to reproduce and ship times the tens of thousands of copies they would likely sell) to create this. I suspect it's more likely intended to be a revenue generator ... and will end up making life more miserable for techies who know what to do (but now have to do more needless paperwork because their PHB bought one of these).

$250 is also excessive to expect to get independent reviews of whether this document is worthwhile. If my PHB brings it up, I want to be able to say it's worthless (if it happens to be) with some foundation for that statement. But without independent review, or being able to review it ahead of time myself, I won't really know. Normally I'm open minded on these things, but for now I'll have to default to it being a worthless document.

Re:Not this again ...

[susano_otter]

How weird...

First you make some assumptions about the contents of the document. Then you say you can't judge the document until you've read it. Then you say that you must conclude that it is worthless.

None of that makes sense. You haven't seen the document, so why should we accept your assumptions about its contents? You say that you can't judge it until you've seen it, but then you judge it anyway. You say it must be worthless, but the only basis for this is the contents, which you say you have not seen.

Wouldn't it make more sense to judge this mystery document based on the performance of past documents written by the same organization?

Has this organization screwed you on the CAT standards? Were interested parties justified in buying a copy of previous standard specifications by this organization, and complying with those standards?

If you're going to reject this standard, shouldn't you at least reject it on the basis of what you know, instead of rejecting it on the basis of things you admit you have no clue about?

Isn't it kind of embarassing to say, "I know nothing about this document, and until I do, I can't judge it, but in the meantime, I'll go ahead and judge it anyway?"

Re:Not this again ...

[Skapare]

You've got it mixed up.

I'm forced to make assumptions about the document. Those assumptions are going to be coming from what other people say, after the presumably read it. Unfortunately, I won't be able to tell whether they did or not. Still, it does appear to be mostly negative from those who discuss enough particulars to indicate that they know what it covers. Keep in mind I weigh the judgement of a techie far above the judgement of a business manager (e.g. PHB) type, because the former are almost always right, and the latter are almost always wrong, when it comes to anything about computers. FYI, I know there are exceptions in both groups; I've met some.

As for judging the document on the past of the organization, I cannot agree. The scope of a standard for a cable and the scope of a standard for a data center are worlds apart. The CAT 5 standard is more about manufacturing something that essentially needs to be substitutable from any vendor. I've designed 4 data centers, and none of them were anywhere like any of the others. I didn't touch the cooling system designs on any, but I did the electrical power parts on 2 of them and there I did find a very reasonable document called the National Electrical Code, which at over 700 pages is a much more substantial document than anything you can put in 148 pages. And yet the network and computer configurations were far more complex aspects of the data center design than the electrical power was (though the latter could easily kill since the voltage coming in was 480 volts in one case and 600 volts in the other, both three phase).

So I ask, what the hell can a mere 148 pages cover that would be worthwhile to me?

One big problem is lots of groups do publish lots of documents on how to do things, charge lots of money for them, and they really are worthless for actually doing anything other than giving a PHB a warm and fuzzy feeling that in the end wasn't warranted. Multiply $250 times all the different documents being peddled, and that's a lot of money wasted.

If this document is an exception, and is well worth the money, then I think those who are producing it need to take steps to make their document stand above the noise. They need to convince me theirs is worth it. And I won't fall for the argument that $250 is not too much to just check it out, because that argument would also have to be applied for so many others, too. Since I already know from experience that most documents are worthless, I need something more than just an announcement to warrant spending the $250 for it, especially when other geeks are mostly shouting it down.

As for CAT 5 (and what preceeded it), yeah, I really would have designed it differently than it was designed. But, since the design wasn't going to change, I simply dealt with the mess and moved on. FYI, what I would have made different was the pinouts ... which would have needed to be done before CAT 5, anyway.

The three steps

[apenzott]

1. Create and publish an industry trade standard (real or imagined.)
2. Charge $250 per hardcopy or .pdf copy of standard.
3. Profit.

Re:The three steps

[Anonymous Coward]

I think it's four steps. You forgot the bit about devising a way to get people to pay for the damn thing :)

Re:The three steps

[saintp]

Don't worry, it's available in PDF. I predict it's available over eDonkey or BitTorrent within the week.

Re:The three steps

[jd0g85]

1. Buy the $250 standard
2. Condense it down into a book
3. Sell it for only $40
4. Profit

Re:The three steps

[Bogtha]

That's common for standards organisations. You have to pay for things like Unicode and SGML as well (if you want to write a proper implementation and not just wing it, that is).

Doomed to failure?

[Toby The Economist]

Be interesting to see if it's useful.

When you specify something like a cable, it's straightfoward to get it right, because the job the cable does and the way it's used is very well understood and doesn't vary between users.

With something complex like a data center, there's so much variance in how they're operated, exactly what they do, where they are, etc...having a standard may well *not* fit everyone's needs, either because their needs were not perceived or understood at the time or because their needs simply cannot be met by the standard.

--
Toby

Re:Doomed to failure?

[JanneM]

With something complex like a data center, there's so much variance in how they're operated, exactly what they do, where they are, etc...having a standard may well *not* fit everyone's needs, either because their needs were not perceived or understood at the time or because their needs simply cannot be met by the standard.

True.

But when you have a formal standard, you have something to measure against. Every aspect of the data center design is not only standardized, but the how's, why's and therefore's are spelled out. If you suspect the standard doesn't meet your needs in some respect (a clear lack of surround sound for late-night fps tournaments, say), it makes it clear exactly how your criteria changes the requirements, and it makes it much easier to see how it could impact the rest of the design.

So even if you use not one single recommendation (we need the disco ball, damnit!), you have something reasonable and well documented to compare against, which makes your job easier.

Re:Doomed to failure?

[duffbeer703]

Large companies or gov't agencies already have standardized data center processes. Go into a cell phone tower equipment room, and you'll find that one room is practically identical to the next in most ways.

When AT&T built its long-distance microwave network, every center was the same as well. If you're running a nationwide network of branch data centers, you need standardization.

Re:Doomed to failure?

[poot_rootbeer]

So even if you use not one single recommendation (we need the disco ball, damnit!), you have something reasonable and well documented to compare against, which makes your job easier.

Until it comes time to justify to the PHBs and bean-counters why you didn't follow every recommendation in the standard, to the letter. Sure, YOU know that your custom solution is more appropriate than the baseline, but how are you going to defend it to people that don't understand it as well?

Pity it's $250 for a peek

[darnok]

I'm curious to see what this document contains: whether it's an ITIL-like view of the world (e.g. a data centre runs on change management, capacity management, problem management, ...), a hardware based view (e.g. a data centre needs a raised floor to duct cables, air conditioning, secure access, racks, ...) or something else.

Just not curious enough to pay the price to find out

Re:Pity it's $250 for a peek

[Gothmolly]

God I hope it has nothing to do with ITIL, to ensure that people actually take it seriously. On this side of the pond, What Works is usually given more credence than What The Committee Decided.

Re:Pity it's $250 for a peek

[mstansberry]

I've got the table of contents if you want a look. It's pretty much hardware based and heavy on the facilities side. Drop me an e-mail and I'll send you the .pdf TIA sent me.

/. effect

[OneArmedMan]

I wonder if they considered defences to the /. effect when writing this.

Re:/. effect

[Anonymous Coward]

You don't need a defence to the 'slashdot effect'. It's not very powerful at all.

I hope it specifies floor monkeys have unix skills

[Neil Blender]

Man, I am tired of those reboot monkeys not knowing jack shit about unix.

Steal the F*ing Manual

[fyoder]

Just not curious enough to pay the price to find out

Seriously. If all manuals were that expensive there would have been no 'RTFM'. It would have been 'STFM'.

But of course

[frovingslosh]

Sure they charge $250 to look at it. If they just released it then people whould show what foolishness it is and all the flaws in it. But if you pay $250 to look at it, do you want to admit that you paid $250 for a bogus standard? Do you want to explain to your boss why you spent $250 of the company's money to do that? If you sell your time as a consultant do you want to tell your customers you are someone who was duped into paying $250 to see 148 pages of bogus standards, or do you want to paint yourself as someone who has knowledge that they will have to pay to have acces to?

It's the same with hazing or initations, the people who go through it are a lot more inclined to accept the end result, because it's one of the ways they justify to themselves what they went through. I doubt that many will pay $250 to read this spec and then be able to give a completely unbiased view of it.

Re:But of course

[Anonymous Crowhead]

Not sure about you, but if I was setting up a data center, I'd fork the $250. It'd probably be your smallest expense and is bound to have some good ideas in it.

Re:But of course

[frovingslosh]

Not sure about you, but if I was setting up a data center, I'd fork the $250. It'd probably be your smallest expense and is bound to have some good ideas in it.

Well that sure is an interesting comment. Just how many marks do you think that you can round up who would each pay $250 a pop for a 148 page report because it "is bound to have some good ideas in it"? At that price it wouldn't take too many people to convince me to crank out 150 pages of opnion, particular when the subject is as broad as "every single aspect of a data center". Maybe even sell it at the bargan price of $229.99 and undercut this report (but be close enough in price that some might think mine and the one discussed here are the same). And it wouldn't be like I was stealing their money, would it? It's bound to have some good ideas in it.

Re:But of course

[Anonymous Crowhead]

Uh, maybe you didn't notice, but this is about setting up a data center. To me, that implies setting up a facility and an infrastructure, not a rack in closet. To me, that implies spending hundreds of thousands of dollars. To me, that implies the possibilities of paying consultants that much for an hour or two of advice before paying contractors a hundred times that to install your electrical. To me, that does not imply buying a cheap rack, getting a T1 and a router, and sticking 5 servers in whatever closet is available in your office space.

Re:But of course

[Anonymous Crowhead]

and by "paying consultants that much", I meant $250 as in $250/hr, not hundreds of thousands of dollars.

Re:But of course

[budgenator]

Having something in black and white, printed by a whoopdy-do sounding organisation can easily save you that much when some moron electrical inspector comes around pissing and moaning about code violations because your subcontractors are from out of town and don't belong to the same union the the inspector belonged to.

Re:But of course

[dtfinch]

To me, that does not imply buying a cheap rack, getting a T1 and a router, and sticking 5 servers in whatever closet is available in your office space.

You just described our setup almost exactly.

Re:But of course

[Anomylous Howard]

A Typical PHB will buy it in a second, just to have something to wave at his contractors. He'll read the executive summary and skim a few choice paragraphs near the prettier diagrams so he can baffle his PHB with bull. He'll add a few words and phrases to his vocabulary and believe that it was the best $250 his company spent since buying him that unused health club membership.
Yes, I am feeling a bit cynical this morning. Why do you ask?

Re:But of course

[twiddlingbits]

You could undercut the costs, but good luck selling it. If they won't pay the $250 for the data from the TIA which is a recognized and highly repected standards body, why would the pay $230 for it from a NOBODY? TIA has a "brand", you don't have anything to compare to that except maybe experience.

Re:But of course

[temojen]

You guys get a rack? All we got was a wooden shelf.

(I think sometimes all these datacentre folks forget that there are a lot more small businesses out there than the handful of transnationals that can afford thousand machine datacentres.)

Re:But of course

[zevans]

If you sell your time as a consultant do you want to tell your customers you are someone who was duped into paying $250 to see 148 pages of bogus standards, or do you want to paint yourself as someone who has knowledge that they will have to pay to have acces to?

And of course if you're a consultant you'll charge $250 for the time it takes to have above discussion.

A flippant way of making a serious point - $250 is NOT EXPENSIVE for something that's valuable as a troubleshooting and planning aid to anyone.

Re:But of course

[dtfinch]

The rack is very small and only used for switches.

Re:But of course

[Anonymous Coward]

1. That's our setup, too, so it must be some kind of standard.
2. I wonder what would happen if I took his post, copied it into a PDF, and sold it to a bunch of wankers on Slashdot for $250?
3. Profit!

Re:But of course

[Macgrrl]

$250 for some good ideas may not be a rip off. I'm a business process consultant for a large multi-national, my daily chargeout rate is $1,750. That means it's worth a little more than an hour of my time.

Hum?

[Ecko7889]

Table of Contents: Page 1-147: The proper procedure to screw in a light bulb Page 148: Apendix

Re:Hum?

[Samari711]

[this post intentionally left blank]

Re:Hum?

[Anonymous Coward]

[this reply intentionally left blank]

Re:Hum?

[Anonymous Coward]

That's the dumbest thing I never heard.

Re:Hum?

[lcsjk]

What was?

Cheap datacentres

[lamasquerade]

I welcome this standard with open arms. I look forward to the not-too-distant day when I will be able to buy 100m(sq) of Standard Data Centre on eBay for $25. No more un-backed-up un-RAIDed hard drives for my mp3s!

The three steps-To my basement.

[Anonymous Coward]

And you're JUST realizing this? Just what does happen in your basement?

Definition of a STANDARD

[Anonymous Coward]

Ok.. everyone do it MY WAY.

That Webster guy did the same thing when he wrote his Non-ENGLISH dictionary.

Link to the document

[jurt1235]

Does anybody have a link to the document yet, since $1.6 per page of bs is a bit too much. What they did is essentially take all or part of the other standards which are around right now, make a crosssection and dump it in a new document. Sort of standard history or literature thesis writing, no real research, nothing new to tell, just a new cover and 2 weeks at the typewriter.

Re:Link to the document

[Anonymous Crowhead]

Does anybody have a link to the document yet, since $1.6 per page of bs is a bit too much.

Slashdot moral concept #7: If item is perceived to suck, stealing - oops sorry, forgot Slashdot moral concept #6 - infringing it is allowed.

Example: "If $band would put out better songs, maybe I'd buy their album. Until then, I will continue to use BitTorrent to get their material."

Re:Link to the document

[jurt1235]

I will not keep the document, I will not even remember it, I will just glance at it.

Or everything you can eat before checkout in a supermarket is free, however in a restaurant the rules are different again.

Re:Link to the document

[Travoltus]

This analogy will fail when this new data center standard becomes a matter of enforcement, like those building codes that are created partially by taxpayer money but you have to pay a corporation a big pile of money just to look at them.

Corporate welfare at its finest.

Of course, these data center standards haven't reached that point, but who's to say they won't? It's being positioned that way. We'll see whether or not it ends up that way.

Re:Link to the document

[hcdejong]

when this new data center standard becomes a matter of enforcement

Which won't be anytime soon, because "TIA is accredited by the American National Standards Institute (ANSI) to develop voluntary industry standards for a wide variety of telecommunications products." (from the TIA website). They don't have the power to write laws.

Re:Link to the document

[jrockway]

Sample scenario:

Hi, your datacenter isn't TIA-compliant. We won't sell you xyz hardware. We won't sell you xyz fire-suppression system.

Hi, I notice your datacenter doesn't have a fire-suppression system. I have to close it, by law, until it's installed.

Hi, I can't install a fire-suppression system until you bring the datacenter up to TIA-standards.

Needless to say, TIA doesn't have to make their spec law for it to be able to screw your datacenter over.

Re:Link to the document

[ColdGrits]

"Hi, your datacenter isn't TIA-compliant. We won't sell you xyz hardware. We won't sell you xyz fire-suppression system."

Yeah, like that's going to happen! A company refusing to make money by selling you stuff? No chance. Some company will always be happy to sell it to you. Heck, if you can show me that all companies will act in that way, then you've just pointed me in the direction of my niche market where I can make shedloads of money!

Re:Link to the document

[Analogy Man]

Sure they won't REFUSE to sell you equipment, but they could have it void warranty, guarantee or violate lease terms. I think this is a good thing. Suppose you are HP leasing a half million dollar pair of DB servers and the bozo's deploying them have inadequate HVAC...the server room spikes to 95 degrees every time the sun thinks of shining and the servers shut themselves down based on thermal protection on CPU's. Would you guarantee the uptime of those servers? How thrilled would you be replacing fried components on them? Would they be reliable machines when they come off lease?

Re:Link to the document

[ngyahloon]

Not exactly link to the document but here's a pdf of the tutorial for this standard (sort of like an overview of what the specification covers, IMHO)
http://grouper.ieee.org/groups/802/802_tutorials/j uly03/tutorial_1_0703.pdf

Re:Link to the document

[gregjmartin]

You forgot one.
Hi, My name is Elvis. I seem to have gotten lost. Can you steer me towards Las Vegas?

ISO standards for business processing and security have existed for years and while a customer may require compliance as a condition for employment, I've not seen the gov't or a vendor require my participation.

But it is a lot more fun to jump on the conspiracy idea.

Re:Link to the document

[jrockway]

> Yeah, like that's going to happen! A company refusing to make money by selling you stuff? No chance.

Try buying alcohol underage. It boggles the mind... Hello? I want to give you money! Money! How can you say no? Even worse, how did the alcohol industry let the government legislate a reduction in its sales? I thought the government was supposed to look out for any and all corporate interests!

Re:Link to the document

[ColdGrits]

Apart from the fact that your "argument" is completely irrelevant (hint - it is not ILLEGAL to build a datacentre without adhering to an industry-recommended standard...), are you SERIOUSLY stating that nobody underage has ever or will ever succeed in buying alcohol? Because if you are, then boy is the real world going to be a shock for you! I guess you think that nobody has ever managed to buy "illegal" drugs either, right?

PDF HAS BEEN LEAKED CONTENTS FOLLOW

[gutterandthestars]

A data center shall consist of hardware, appropriate HVAC, and gigabytes of pornography.

Right answer, wrong question

[seanadams.com]

The question was how exactly does varying the twist rate help? (See uncle post for a starting point).

Re:Right answer, wrong question

[Anonymous Coward]

To "detune" the coils. If the twists were equidistant (especially in the straighter parts) it would resemble a tuned antenna.

great news for colocation customers

[pokka]

This is great news for people who host servers in colocation facilities.

If you've ever tried to find a place to host your server in the past, you've probably found that not only does the price wildly fluctuate between hosts for no apparent reason, but also it's very difficult to determine exactly what you're getting, even if you take the time and effort to actually visit the site.

I think that the disorganized fashion of colo services allows people to charge ridiculous prices
and get away with things that they wouldn't be able to do in a more stable competitive environment (like charging ridiculous amounts for bandwidth overage and support).

With some sort of standard in place, vendors will be forced to compete on more even ground, prices will be more reasonable, and users won't be afraid to leave their current colo provider because the next one could potentially be even worse.. Not that it will be perfect, of course - just somewhat better.

Re:great news for colocation customers

[Anonymous Crowhead]

I don't think this is really for that, but some sort of manual or book about choosing a colo would be a pretty good idea. Your point about overage is a good one - some companies have outrageous 95th percentile billing schemes, say charging you for the max megabit for the entire month, or just pro-rating you. The difference between the two can be thousands of dollars. Support too - $300 an hour with a half hour minimum can be a lot when you start swearing to yourself that they are sending the security guards out to your rack when you call for hands on support.

Re:great news for colocation customers

[RajivSLK]

"Carrier neutral" facilites in large buildings are your friend. For example the westin building in seattle has many network providers that can be pulled into our cage independant of the company that provides the colo. This forces them to be price competitive as they don't hold our equipment hostage. This is why we are moving a significat number of servers from the harbour centre in Vancouver BC (which is not neutral).

If you want the best price you have to know what that price is and ask for it. They'll say no. Then ask again and again. Then every six months call them up and ask for a better price.

Re:great news for colocation customers

[pokka]

Yep, I think the problem with colocation is that there are so many different things to worry about at once, and that's how you can be taken advantage of - perhaps a data center standard would consolidate 50% of those factors into a single rating so that I can focus on the things that *won't* be part of it, like pricing, hours, availability, etc.

Re:great news for colocation customers

[Jeff DeMaagd]

If it really would lead to much less expensive colocation, why would the operators switch to a new standard so they can charge less?

Re:great news for colocation customers

[pokka]

They wouldn't have a choice. That's the beauty of *fair* competition. Their customers would either demand lower prices, or leave and go with a colocation provider who has them. The reason people don't do this right now is because "colocation" really has no standards - it could be some guy's basement or a deluxe, cardkey controlled secure building. It takes so much effort to find out if the provider you found is a good one, that it's usually cheaper to stick to who you have and not waste your time. But if you could call up a provider and ask "Are you class A data center according to the TIA standard?" you could easily and quickly weed out the ones you don't feel meet the standard of quality. And if *everyone* can do this as well, colo prices would be more competitive.

Alternatives to THIS book

[xmundt]

Greetings and Salutations.
          Interestingly enough, a quick google search for "data center design" comes up with more hits than one can shake a stick at, ranging from free to fairly inexpensive (under $100.00). I have to admit that I wonder if THIS magnum opus has anything in it that these OTHER resources do not cover.

          It never ceases to amaze me at the number of books out there that are supposed to be useful learning tools that are nothing more than a slightly changed rehash of the man pages for a given program.

          Regards
          Dave Mundt

Re:Alternatives to THIS book

[oneiros27]

For a quick overview, there's the book from Sun, "Enterprise Data Center Design and Method", which covers concepts on dealing with airflow, rack placement, power requirements, etc. The first chapter is available from Sun Blueprints:

Data Center Design Philiosophy

Sun also has some articles on disaster planning and such in the Data Center section of Sun Blueprints.

Oh ... and $250 is not a lot of money, when you're dealing with a buildout of millions of dollars, just for the data center (ie, not the actual systems to go into the data center)

148 pages?

[Anonymous Coward]

Unfortunately unable to look at the details right now, but 148 pages is very modest for a current day standard.

If only some of the other bodies would take that as an example. If you have ever had to implement any standard you know what I'm talking about.

Raised Floors?

[Anti-Trend]

I can't afford the $250 at present, but I wonder if they finally did away with raised floors. It wasn't too bad of idea around 40 years ago, but we've got cool modular racks now that make that concept moot, at least IMHO. Plus raised floors look weird, are fairly expensive to implement (especially for smaller firms with little cash), and get really nasty under there over time. Besides, telco has done without that design for quite a long time, seems to have worked out fine for them.

-AT

Re:Raised Floors?

[DaEMoN128]

After making my living as an installer... raised floors are a GOD SEND!! They are a pain, but they look 1000 times better than cable tray or ladder rack with 2000 cables in them. They make routing cables easier. It is no fun setting up some rig in order to hang cable tray 20 ft. off the ground. You dont want ladder rack the whole way... nothing wrong with it, but why when you can use a cable trough with a cover that allows the use of a pull tape. Raised floors help with cooling (forced air from under the cabnet). Also, they dont have to get nasty under there.... just make sure your standards for presentation are clearly stated in the contracts and enforce them.

Re:Raised Floors?

[Anonymous Crowhead]

I wonder if they finally did away with raised floors.

At my colo, they run cold air from the HVAC under the raised floors and suck it up through the cabinets with a fans (and presure from the HVAC). They cool the cabinets, not the entire facility. It's odd being in a colo that is warm after freezing my ass off in previous facilities, but temperature monitering equipment of ours tells us that the closed cabinets stay quite cool.

Re:Raised Floors?

[Linker3000]

You would have loved the computer room I did some work in during the 1980s - the builders got the dimensions wrong and there was a 4 foot void under the tiles instead of 4 inches!! The owners had to have special pedestals made to hold up the tiles and they actually put some servers under the floor!

Re:Raised Floors?

[wirefarm]

Did they later convert that space to a cubicle farm?
I think I worked there.

Re:Raised Floors?

[marquis-cablewitch]

But without raised floors where am I supposed to hide the bodies?

Re:Raised Floors?

[Jon Chatow]

The data centre at CERN has a 1 or 2 metre-high void, IIRC. But then, they have rather a lot of cables. It was a deliberate decision.

Re:Raised Floors?

[Splab]

We got the same thing on our new site, complete with a couple of v8 diesel engines for backup power and 3 cooling units the size of my apartment.
Looks very impressive :)

Re:Raised Floors?

[Chanc_Gorkon]

Raised floors are definitely something you need. What we did was run the backbone cable to patch panels in the server rows. The cabling to the servers are overhead racks with troughs for fiber. This was a huge improvement to the old room. We pretty much bought a switch and prewired every panel. Now when we need to add a server, we have out cisco guy add it to the switch config(we also give him the MAC address at this time too), they tell us which jack to use and we get our wire expert to run the cable from the patch panel to the server. It takes far less time to get new equipment wired becasue we don't run a cable from the equipment to the switch. The only hard part is power, and even that isn't hard. Our Pseries racks all take the same power supply so we had twist locks placed at strategic locations. When we get a new server, if we need another power supply, we order it to and plug it into a empty twist lock receptical. Standardization of Data Centers is sorely needed. The types of standard don't mean every center will look the same, but it will mean that all equipment in the future will have the same or similar requirements for network drops and power. This makes data center planning MUCH easier.

Re:Raised Floors?

[spudnic]

You must not work in a large data center.

I can't imagine what our server farm would look like if we didn't have a raised floor for routing cables, fibre, air, etc. Do you have any idea how much heat a rack of 1U servers can put off?

Re:Raised Floors?

[tengwar]

I work for a telco. I've never seen a machine room without a raised floor.

Re:Raised Floors?

[Hasai]

"I wonder if they finally did away with raised floors."

Are you out of your mind??? I don't know what sort of rinky-dink operation you're used to, but without raised floors our server rooms would rapidly turn into a bloody jungle and a hazard to both life and limb. :P

Re:Raised Floors?

[tgd]

You sure about that? I've never seen a "real" data center with just four inches below the floor. I think the least I've ever seen was maybe 3 feet, and I've been in several you could comfortably stand under.

The bigger the space, the more serious the data center. I did six months of consulting working in a data center that had a six foot raised floor, which wasn't enough to walk easily under because of the ducting and cable trays, but made moving around and finding stuff a lot easier. They had great filtration, too... it was spotless under there, and ironically a lot warmer, so you'd occasionally see people sitting under the floor.

Re:Raised Floors?

[Anonymous Coward]

A previous call center I worked at hosted rats and vermin underneath their raised floors :)

Re:Raised Floors?

[GnuTzu]

Some telco offices don't have raised floor. Instead, cables are run in overhead trays. You want, and I mean really want, those trays to be strong and securly mounted.

Re:Raised Floors?

[bhsx]

No, they made it the 7 1/2 floor. It's a real bitch stopping the elevator there; but it has some wierd perks.

Re:Raised Floors?

[Fulcrum of Evil]

But without raised floors where am I supposed to hide the bodies?

Even with raised floors, you eventually run out of space. This is why every sysadmin worth his salt has a lime pit.

Re:Raised Floors?

[marquis-cablewitch]

I'm trying to think of reasons to convince $EMPLOYER that they need to get me a small plot of land with a big hole in now...

Surely a better and more ecologically sound method is just to own a large wooded area and bury people under the trees? Although it would I admit take longer.

I like cake.

[drinkmorejava]

I look forward to the day when data center security is standardized and one hacker can bring down the whole world.

Re:I hope it specifies floor monkeys have unix ski

[DenDave]

LOL!!

This would be neat! Of course I doubt that the CAT would dare to take such a leap... I like your term reboot monkey, I think I'll use that... I usually refer to them as clickologists...

On another note, did any of y'all get a patch for your ciscos yesterday? It seems to be a pretty big one...

you know what that means

[commodoresloat]

Less crosstalk, more crossrock!! Yeah, Stryper!

Re:you know what that means

[Anonymous Coward]

but i thought the twisted sisters said you can't stop crosstalk?

racks/cabinets

[Anonymous Coward]

they better spell out in clear terms that round-hole racks are concentrated evil. square-hole racks only.

it's the 21st century. if i have to assemble a goddam set of rails and get a power screwdriver out to install a pile of servers again, i'm gonna kill someone.

square hole racks support snap-in rail sets for tool-less install, as well as nuts for old school rails, and are worth their weight in gold.

since the parent is a slashvertisement anyway ($250 for a copy of said document) i'll go ahead and plug netapp, since i just set up a shitload of gear recently using their cabinets (no i don't work for netapp)

i'm pretty sure they just slap their logo on some 3rd party cabinets, not sure who ultimately makes them, but they rock.

anyway, it would be interesting to know if they come to the same conclusion in this doc. if they don't, it's worthless.

Re:racks/cabinets

[Hasai]

HP makes some pretty spiffy square-hole racks, as well; I was a little leery of the design at first, but now I don't know how I lived without them. Snap-in mounting rails: Whee!

I'd like to have a little heart-to-heart with HP's shipping department, though; the last rack they sent me arrived bent. It took me and another guy three hours with a 10-ton hydraulic ram to get the thing back into shape. What the hell did they do to it, run over it with a locomotive?

:(

Re:racks/cabinets

[altek]

Umm, why not just make them send a replacement?? I know I wouldnt bother with 3 hours and heavy equipment when I can just get them to reship. I mean those things run like $1500, its not like they wont accomodate...

Re:racks/cabinets

[Anonymous Coward]

Somebody probably backed a fork lift into it or dropped it off a loading dock.

Re:racks/cabinets

[Anonymous Coward]

They could specify the depth for racks too. Naive me thought 19" was some kind of standard and bought a bunch, that I now have to modify becuase the long server slides are one inch to long (with the movable connection in the shortest position) and the short ones are one inch too short (with the movable connection in the longest position). Luckily there are square hole spares to buy I can mount one inch in front of the back of the rack.

Link to the document-"PDFs want to be free"

[Anonymous Coward]

"...recently published a 148 page document meant to standardize the design considerations for every single aspect of a data center.[emphasis mine]"

"This analogy will fail when this new data center standard becomes a matter of enforcement, like those building codes that are created partially by taxpayer money but you have to pay a corporation a big pile of money just to look at them.[emphasis mine]"

The only "you" that's going to be spending money on the PDF is your boss. Otherwise don't worry about a "link to the document".

Finially a DC standard

[jimmypw]

OMG my dreams have been answered. Finially the method of putting in a rack nut had been standatdised.
The method of putting in a patch cable is now an Easy-to-follow, step-by-step process.

I'm all for standards, but not the ones that cost $250 (£138.61) and probably wont change what you already do.

Re:I hope it specifies floor monkeys have unix ski

[Anonymous Crowhead]

My two favs:

#1. Driving all the way to the data center at 4 am to find the keyboard plugged into the mouse port - "The keyboard's not responding", says the "engineer".
#2. Driving all the way at to the data center at 3am to see /sbin/fsck-y on the console - "It says command not found", says the "engineer".

Installed X and AIM after that and made the "engineers" read the commands off that provided they could get that working.

AMEN

[DaEMoN128]

I agree completely. The "industry standards" for many things are used alot, even if they dont partularly apply to your site. I would love to see people dress the cables in the back of a cabnet with uniform bending radius' and with a proper service loop. I would love to see that in every data center I work in. I wouldn't have to replace cables when a channel bank gets moved. I would love to see this fiber I have to deal with be secured properly and not cinched down hard enough to break your 1st. 6th, 18th, 32nd, and 72nd strand. That would save me a lot of work. I am not going to pay to see what they say... but I bet its just a culmination of proper install techniques and "industry standard" dressings to make the tech's and the installers times easier. Remeber "if it aint pretty, it dont work", err... that's what they told me in BICC anyways.

Re:I hope it specifies floor monkeys have unix ski

[rasjani]

Man, unix skills ? I'd love if they knew what is the difference of dumb-terminal and real monitor - and why they cannot use one with the another..

Already pretty standard, I think.

[Lellor]

My experiences with data centers in Canada, the US, the UK, and South Africa so far lead me to believe that they are already pretty standardized - there may be slight differences, but all in all it's pretty amazing to see how similar data centers are when one is visiting countries that vary so greatly in other respects.

I guess that could be advice for immigrants from other countries who are in the tech industry: if you feel homesick, just spend more time in the data center - it will very likely look just like the data centers back home.

This is a really bad idea, sort of

[DaEMoN128]

From TFA, this is a "checklist" for CIO's. Last thing I need is my PHB having a list to check off and thinking they are requirements instead of suggestions. You never give a PHB this much info. They dont know what you are doing a good part of the time anyways... and a little info is more dangerous than none.

It can be a good idea if the techs get a hold of it though and stop giving my 2 inches of slack on these fiber runs and give be a proper service loop with good cable dressing instead of the rats nests I've had to fix recently.

Re:This is a really bad idea, sort of

[ch-chuck]

A saying once was that a good advisor finds out what his boss wants to do, and advises him to do it. I've noticed a lot of PHB's just ask their crew what they think needs to be done, then tells them to do it.

And *never* "think out loud" to a PHB, as in enummerating a list of possibilities to check into - his little mind will just grasp onto the first one and decide that that's *it*, whether it actually is or not.

Re:This is a really bad idea, sort of

[Anonymous Coward]

Last thing I need is my PHB having a list to check off and thinking they are requirements instead of suggestions. You never give a PHB this much info.

But you do...here's how I would approach it;

1. Look over the standard and understand it.

2. Identify any gaps between the standard and your site.

3. Push to have all the standards in the spec implemented, and put costs behind each one.

4. The CIO balks and refuses to pay for the changes.

5. Everything returns to normal, plus your CIO will break into hives if the standard is even mentioned again.

Ah, the joys of manipulation!

Re:This is a really bad idea, sort of

[JorDan Clock]

...good cable dressing...

I prefer a nice Italian for my cabling, but with the higher standards of CAT6 only the finest you can find is suitable, such as a very expensive red wine.

Got a torrent?

[Anonymous Coward]

*duck*

1 shoe does not fit all

[zenst]

Being unable to lash out and indeed darned if I woudl ever pay $250 for a datacentre for dummies kidna guide given that every case has to be treated on its own merets.

I wonder if they cover aspects like power phase balancing given alot of places have 3-phase and we all know how box's move about so the aspect of auditing the balancing across the 3 phases comes in. Why well power costs given you pay for the highest usage over the 3 and then there is the UPS aspects and resiliance aspects.

Oh and DR sites, if you have enough eggs for one basket then you need backup plans.

As you who work in the feild can imagine there are sooo many aspects that need to be looked at and whilst a guidline/crib sheet is always welcome no two datacenters are the same or some manufactuirer would be making a killing selling them in a box ;0.

The smallest deatil on one site becomes the largest priority on another. Hell i've even set up a site were we had critical servers spread across the area just to give that extra resiliance incase a train derailed and came thru, albeit remote possibility but still was a potential, and all before 9/11 which in itself shows the level of which murphy's law can stoop.

So whilst this type of thing is nice I do erk when people hold it up as a definitive guide as the only people fooled by that are those that dont know and highup managers and we all know a little knowledge is a dangerous thing.

Still its nice to have stuff like this as long as they are viewed in the context of guidlines and not guides in themselves.

I mean if all doctors were to follow a needle injection guidline of insert needle in left arm 13cm down from main index finger and on the horizontal plane of the wrist joint, not everybody would get the injection into a vien as we are all not exactly the same albiet similiar and indeed some people have no index finger nor left arm.

I personaly wish they would invest more time in a cable standard were the outer colouring deformed in relation to the number of times the cable is bent, now that would in itself be a much more useful input into datacenters in that they could pre-emptivly identify cable that is going past its shelf life or been over abused/used and liable to induce errors.

Re:1 shoe does not fit all

[danielrose]

I mean if all doctors were to follow a needle injection guidline of insert needle in left arm 13cm down from main index finger and on the horizontal plane of the wrist joint, not everybody would get the injection into a vien as we are all not exactly the same albiet similiar and indeed some people have no index finger nor left arm.

Sorry, if you have no left finger or arm, I cannot give you the injection, there is just no way to do it.

Re:1 shoe does not fit all

[Anonymous Coward]

Here's an idea: go back to elementary school and learn how to spell and form coherent sentences.

Re:1 shoe does not fit all

[zenst]

lol - was late and well; Think faster than I type, and cuting and pasting into spellcheckers is a pain.

But glad the rest panned out for you :).

April's fool?

[VincenzoRomano]

The document TIA-842 has been issued on April 1, 2005!
I'd not take in too serious consideration this doc!

Re:I hope it specifies floor monkeys have unix ski

[Anonymous Coward]

Appropriate term. Some years ago, I set up a Linux box with Oracle for a bunch of them at a client site that was traditionally all WinDoze and tried to train them...hopeless. It was hilarious to watch as they struggled with the concept of CLI. It was exactly like that: like watching a bunch of monkeys trying to fuck a football.

Pfff..

[Anonymous Coward]

They charge $250 for a Book?

Dude.. and I thought the DVD's and MP3's I'm downloading were overpriced...

step 4

[RMH101]

realise your insurance company is eventually going to require a DC built to whatever standards they force through....

waste of time

[smithcl8]

Who funded this standard? What a waste!

Just standardize the rack widths and the use of the square holes....I'll put it in a room as I see fit.

Raised floors and the like are great, but to "standardize" it is silly. Next thing: standardized IT personnel. This way, companies can interchange their folks with the greatest of ease!

Re:waste of time

[servant]

IT personnel are alrady standardized. They are on the other end of a telephone and all speak proper English with an Indian accent.

Re:waste of time

[Anonymous Coward]

No, they are standardized by stupid things like MSCE, which is meaningless in the real world. It is about what paperwork you have, not what you actually know.

Re:waste of time

[GnuTzu]

The telco industry still manages to run at a higher level of reliability and quality than that part of the IT industry that has chosen to ignore such standards. Standards mean that facilities can be managed in a consistent fasion. Seriously, many problems are really a matter of someone settings things up in an unexpected manner.

Sadly, the cell phone industry as well as the Internet is forcing the old telco way to leave behind some of these assurances. Quality and reliability will suffer everywhere.

There will come a time when someone will light a candle during a power outage. The candle will start a fire. But when they go to call the fire deparment, niether phone nor cell phone are working. Cough cough.

Re:waste of time

[cfuse]

Next thing: standardized IT personnel. This way, companies can interchange their folks with the greatest of ease!

You Indians all look alike to me ...

But of course-Cheapskates.

[Anonymous Coward]

In other words, the majority here are "penny wise, pound foolish" aka cheapskates. Can't imagine why, since it's their boss that will be forking over the money (including the PDF). Not them. Although I'm certain their boss will approve of their "Geico" thinking.

Any Zinc Whisker failures?

[Overzeetop]

Seriously, search /. on it.

Purpose of the Datacenter

[PhYrE2k2]

But when you have a formal standard, you have something to measure against. Every aspect of the data center design is not only standardized, but the how's, why's and therefore's are spelled out.

But not all datacenters are equal for a reason. I've seen maybe 20-30 datacenters in the past few years for various clients and they all have different features, different offerings, and different goals.

I'll list a few of the big differences I've seen in my experiences:

Some want to be in the downtown core, close to many businesses, but charging a premium for the space. Others claim that being in the outskirts of the city provide security in the event of any problems (mainly hyped due to 'terrorist attacks').

Some feel the need for N+2 generators, others more. Some feel that a fallback to city power if their PDUs ever fail is good, and others feel there should be a whole other protected power distribution system (at an extremely high cost for something rarely used).

Some like cooling each rack from the top, others blow it up every other isle and suck air down on the opposite. Some cool the whole room, claiming lots and lots of cooling units around the outside does the trick.

Some like the datacenter two stories underground. Others claim that they're a first target for flooding and other problems stereotypically associated with a basement. Others say that the datacenter on the 10th floor of a tower is inaccessable and subject to other security feats of the building.

Some like dedicated buildings, others like quietly slotting themselves in office towers.

A few I went to were monitored from 3000+km's away, and others had 24/7 onsite staff. Some had technical electronic keys, and others a simple mailbox key. Some had biometrics, others just a key.

One I went to even had outter walls capable of withstanding most missiles. Others had windows with only paper over them for security reasons.

Some let you roam freely by a security personelle and simply log equipment, others weigh you on the way out to make sure that you didn't take anything you didn't show up with without signing it out.

The point is each of these serves a very different purpose. If you are going to have lots of untrusted people working on equipment, it's important to make sure nobody takes anything. Each one has its advantage and disadvantage, and I don't think any one of them is 'right'- it's just trying to find a solution to problems that experience has provided.

Is there a right answer with anything? Who is to say that any answer is right or wrong? They're just different solutions to the problem. If power stays up, systems are secure, systems get cooled, and the network is available, who is to say the solution is wrong?

-M

The best-organized datacenters I've seen...

[guitaristx]

...don't have raised floors. They use cable trays that are hung from the ceiling. It's a beautiful thing if you've ever had to lay on your belly reaching under the floor trying to fish out a cable while the refrigerated air is blasting your face and drying out your eyes and nose. If this spec wants raised floors, I'll save more than $250 for myself and my company by designing the datacenter without this goofy spec, and use with easy-maintenance stuff like overhead cable trays.

Re:The best-organized datacenters I've seen...

[subreality]

Actually, you can use raised floors for another purpose - cooling. Your AC pumps tons of cold air under the floor, and then you have vents under each rack. You use racks without vents up and down the sides, and you end up with a cold wind tunnel blowing right through each of your racks.

The cleanest data center I've seen did this, ran the power under the floor, and ran all the data through overhead cable trays.

Re:The best-organized datacenters I've seen...

[T-Ranger]

This always seemed strange to me. Hot air rises. You pump cold air into the floor? WTF?

Shouldnt you have a false (T-Bar) ceiling, pump cold air into that, preasure plus gravity pushed the cold air down the rack, and you then suck the (now hot) air out the raised floor?

Re:The best-organized datacenters I've seen...

[EndlessNameless]

:::Shouldnt you have a false (T-Bar) ceiling, pump cold air into that, preasure plus gravity pushed the cold air down the rack, and you then suck the (now hot) air out the raised floor?:::

No.

Hot air rises out of the topside ventilation ducts when it is "pushed" upward by the heavier, colder air below. The airflow is simpler that way. You can then cool the air and return it to the lower level elsewhere.

It's a closed loop with servers making the air rise on one side (by heating it) and AC making it fall on the other side (by cooling it). The cooling scheme is one massive convection current.

Yea, UNIX admins are better?

[cbreaker]

I've seen one too many hacked up script, server hanging by a thread, major changes deployed during production hours, condescending attitude bullshit UNIX admins in my time as an IT admin that I have no more respect for them then any Windows admin.

Ohh you can type fsck? Big fucking deal. You're not special. I love the command line and it's an awesome way get things done but it's also an awesome way for a dipshit admin to fuck everything up with little undocumented custom scripts and programs *everywhere.*

You generally don't get into that deep of shit even with the worst of Windows setups. Sure, it's quite easy to fuck it all up, but it's also an order of magnitude easier to clean up after a shitty Windows admin.

I like all stuff, Unix/Linux/Windows/Whatever. A Bad IT person is bad no matter what, and a good one is good whether they're on Windows or Solaris.

Re:Yea, UNIX admins are better?

[Neil Blender]

Get off it dude. I am talking about people who work in data centers whose job it is to go to your server and do stuff for you over the phone. You'd think these people would have some basic knowledge of various OSes. You'd think it would be a job requirement. From my experience, they have absolutly no unix experience, down to the simplest commands and they are of absolutely no fucking help even though the data centers can charge you hundreds of dollars an hour for it. And even if you don't require their help, your still paying for it as some part of your monthly bill.

Re:Yea, UNIX admins are better?

[Thuktun]

Agreed. The same thing goes for most jobs, it's not the type of tools that are used to produce the job that counts, it's mainly the quality of work. (Quality including things like the time it takes, the thoroughness of their work, the durability of the finished product, etc.)

Expensive set of standards

[otomoton]

Do they really think people are gonna pay for this? Especially seeing how easy it is to get these type of things from "the usual sources". It's not going to be like windows where your company can get audited. They aren't going to come in looking at blue prints and go "Hey, did you license this layout correctly, it looks like the one in the standards book, and does not appear to have licensing anywhere.". So what's to stop me, or anyone that's laying out a data center, from just "aquiring" this book and using whatever good ideas it may or may not have.

Re:Expensive set of standards

[csirac]

I'm sure you're right in your skepticism that "people" won't be buying it. And that's because it's not for "people", it's for datacentres. A place where $250 would barely even pay for a weeks (days?) worth of electricity. You're talking about places that have an annual turnover that is well into the 7 figures, at least.

The standard is used as common ground for reference and comparison. It doesn't just outline specs, but justifies them too. So they can look to this document to see how they may improve.

And as for being audited for not paying for a "standard" for which you haven't paid for and have accidentally implemented on your own, this is irrelevant. You can use the contents of just about any standard I know of to your heart's content, this includes USB, Bluetooth, and MIDI - sure, you need to pay for the full spec - but if you build your work around an "abridged" or alternative version of the spec, nobody is going to care.

They only care when you say "My device is Bluetooth!", even in that case you have to do much more than buy the spec (big $$$ on having a testing facility evaluate your "bluetooth" device's implementation so you can get that cool little blue logo on it).

I think the only reason these guys would be upset if you accidentally came up with a datacentre that looked like the one(s) in their spec would be.... actually, I have no idea. I doubt they'd care one jot.

Why we need something like this

[IGnatius T Foobar]

There's currently no "gold standard" for data centers. This is actually a bad thing, because self-important corporate audity types want some sort of "this is ok" label they can slap on a facility (or worse -- one to look for when selecting a facility). Right now they're all using SAS 70 as the standard certification to look for, which is ridiculous because SAS 70 is really more about bean counters and accounting practices than it is about data center facilities. Something actually designed for our industry would be more appropriate.

Whether this is it, is another story.

The Basics....

[Duncan3]

Above sea level. Check.
Somewhere the US/Chinese government won't be monitor ing everything... still looking.

Sweet Zombie FSM!

[Safety Cap]

$250US for a farkin PDF?!

Riiiigggghhhtttt....

BitTorrent Link

[eventDriven]

The publication is available in either hardcopy or .pdf for $250 through IHS Global Engineering Documents.

Where's the BitTorrent link already?

Pirate the F*ing Manual

[Skapare]

Now days, that should be 'PTFM'.

Re:Pirate the F*ing Manual

[rubycodez]

Pirate is such a nasty word, let's P2PTFM

Standard document

[jcdick1]

Having read through the draft of this standard given to me a vendor in preparation for a major infrastructure overhaul, I have to say that this document was a godsend in getting what needed to be done made possible.

This standard isn't for the SMB or small colo facilities. This is more for the big corporate datacenters (my workplace is approx. 100,000 sq ft, and a 2000+ port SAN). These kind of places don't blink an eye at $250 for a book. Of course, in places like this, a vendor would most likely give a freebie copy of it anyway.

This standard doesn't just say "Have raised floors" or "Don't have windows, but have X number of egresses." It says "For X wattage of computing capacity in a given area, you need Y BTUs of cooling capacity with the ability to move Z cubic feet of air per minute." Things of that nature. Its actually a pretty detailed document.

Of course, no standard fits all situations. Probably the only data center that could actually meet these standards to the letter is one that is only just breaking ground. A thirty or forty year old data center that has had its permanent infrastructure (plumbing, etc) worked and reworked and patched and reassembled can be pretty expensive for anyone to rearrange to meet these.

Re:Standard document

[Anonymous Coward]

How much of what is in it is already covered in the TDMM (Telecommunications Distribution Methods Manual) and NIQA (Network Installation Quality Assurance). I mean there have been standards for many things that data centers have ignored over the years like the Bellcore or Telecordia standards for cable tray utilization etc..

This could be usefull in getting needed things implemented by the PHB but just seems like a TIA rehash of already existing guidelines.

Sooo, where is the torrent?

[nurb432]

Seriously though, isnt 250 a bit much for a unproven document?

Perhaps its worth that, but i personally wont take a gamble that it is. Ill just keep doing things that work.

Would have been nice to read it.

Raised floors are passe

[wsanders]

The last few data centers I've worked in have had raised floors for flood protection. Now that positive-pressure raised floors are out of fashion, there are a lot of raised floors out there. Not for Katrina-sized flood protection, but for garden-variety HVAC and water main leaks; you need an alarm system since the area is hidden.

I actually was wondering if it might be good to use the raised floors as exhaust plenums. You could suck hot air near the ceiling in via raised stacks. The space was never good for anything except collecting shards of wire and starting fires.

The newest data centers I've been in don't even have raised floors at all.

Re:I hope it specifies floor monkeys have unix ski

[gregjmartin]

And in a perfect world, application folks would write code that understood that not all data is on DASD, but might actually traverse a network to get to the app; that network connections in the real world drop occasionally and your app should deal with it; that the admin caring for your server is a human being who, whether you believe it or not, is trying to do the right thing and was not born with a full and perfect understanding of all things - as you, no doubt, were.

an AD or an article?!?

[Anonymous Coward]

FTA "The publication is available in either hardcopy or .pdf for $250 through XXX Global Engineering Documents."

What the hell? Is slashdot being used for slashvertising? If thats the only way it is being made available why is it an article and not an ad?

MECC (OT)

[Seraphim_72]

MECC? Did you get your nick from the old Minnesota Educational Computer Consortium?

Curious

Seraphim

Standards

[torrents]

What good are standards if companies don't manufacture their products to spec... Many organizations also have their own internal guidelines for things such as hardware and site layout... It's hard to see how a one size fits all approach can be applied to something as unique as a datacenter...

not exactly thorough

[Morgant]

Right now I'm about half way through Cisco's "Data Center Fundamentals" book. It is very much an overview and it is 958 pages. While Cisco may not exactly be unbiased, they're not exactly slaps either. What I'm saying here is: if Cisco thinks they can only give me an overview in 950 pages, what can TIA possibly be providing in just 150?

Doomed! Curical missing pieces

[Roadkills-R-Us]

I have it on good authority that some of the most impoartant items are missing:

1) How to determine caffeine / worker ratios, where to put the coffee pot and soda machines, whether and how much to charge, and a list o vendors who still deliver Jolt.
2) Air lock standards so you can crank up a stereo in the machine room loud enough to drown out the machine noise, without irritating fellow workers in cubes, managers walking around, and customers waiting in the lobby with different musical preferences.
3) Minimal standards for protection against Slashdot Effect.

hence the document is about as valuable as most New orleans real estate at the moment.

Re:I hope it specifies floor monkeys have unix ski

[cbreaker]

Not to mention, it's not like the graveyard shift at a datacenter is going to attract those born-with-skillz people anyways.

Some people expect too much.

Sun has an online document

[lobotomy]

Sun has Site Planning Guide for Entry-Level Servers for free.

Chapter 1 describes site planning assistance that is available from Sun, site planning process and concepts, data center location, and route to the data center.

Chapter 2 explains environmental requirements of the data center, including temperature, humidity, cooling, and airflow.

Chapter 3 gives information about rackmounting the servers and how to locate rows of racks in the data center.

Chapter 4 discusses power and cooling issues relating to the servers, including power sources and heat output and cooling requirements.

Chapter 5 lists shipping, physical, configurations, electrical, environmental, rackmounting, and clearance for service specifications for the servers. It also provides specifications for Sun cabinets.

Chapter 6 provides a site planning checklist that you can use when planning your data center and preparing for system installations.

Peer the F*ing Manual

[Skapare]

Even better!

Another Datacenter physical design book

[impeachsarbanes]

I've worked in good datacenters and bad datacenters. As a network engineer my primary pain point is cabling. One datacenter I work in takes 10 minutes per patchcable to install. Another takes 1 minute per patchcable. The difference is how the structured cabling and cable management are provisioned when the datacenter is built. If you are installing a datacenter it is worth spending some money to make certain the people making decisions don't screw up, resulting in pain for you when installing patchcables next year. So I support this $250 document and hope my management will buy it (a bit spendy to buy just for general reading). I highly recommend "Build the Best Data Center Facility for Your Business" by Douglas Alger, published by Cisco Press. 375 pages. $55. A bunch of lessons learned from Cisco's datacenter operations team (I did shell out my own money to buy that one for general reading). Full disclosure: I used to work for Cisco and slightly know the author.