Xiotech Unveils Disruptive Storage Technology

Lxy writes "After Xiotech purchased Seagate's Advanced Storage Architecture division, rumors circulated around what they were planning to do with their next-generation SAN. Today at Storage Network World, Xiotech answered the question. The result is quite impressive, a SAN that can practically heal itself, as well as prevent common failures. There's already hype in the media, with much more to come. The official announcement is on Xiotech's site."

Comments

Looks like it's next-gen primary disk...

[Xaedalus]

Okay, so at a brief glance it's looking like next-gen primary disk storage. I didn't see any mention of which RAID it is (although I'm thinking they're probably going RAID 10??? Maybe 6?). What's cool though (at least by my opinion) is that it's going to cut down on SAN errors through self-diagnosis. Interesting, will have to check through the white paper.

Re:Looks like it's next-gen primary disk...

[maharg]

not only self-diagnosis, but onboard disk remanufacturing ffs

100 or so engineers involved in the project have replicated Seagate's own processes for drive telemetry monitoring and error detection -- and drive re-manufacturing -- in firmware on the Linux-based ISE. ISE automatically performs preventive and remedial processes. It can reset disks, power cycle disks, implement head-sparing operations, recalibrate and optimize servos and heads, perform reformats on operating drives, and rewrite entire media surfaces if needed. Everything that Seagate would do if you returned a drive for service.

Re:Looks like it's next-gen primary disk...

[hackus]

Exactly what we do not need.

Next generation hardware that is patent encumbered and will require a lawyer and several court proceedings for anyone and everyone to get their data back.

I mean come on, when is the industry going to figure out we do not need proprietary, closed storage solutions that are a rehash of the old IBM AS/400 days when you could only buy super expensive IBM gear.

No thanks I will take my open code and commodity hardware and build solutions that will kick this patented solutions arse at 1/100th the cost.

Besides, if these features are really worth their salt the open source community will provide them sooner or later. Preferably in Europe where these silly patent claims that this product is so unique nobody could possibly figure it out, gimme a lot of money because I am brilliant.

Not brilliant and not worth the cost in my opinion. (Both in restrictions due to the patents and infrastructure choices this product imposes and the cost in currency).

-Hack

Re:Looks like it's next-gen primary disk...

[Anonymous Coward]

Please supply code for clustered storage controllers with no single points of failure. Also, please provide code for synchronous and asynchronous disaster recovery replication. While you're at it, scale said commodity system to > 1PB.

Oh, and I'll want 24x7 support for this system and all its components with 1 hour onsite service.

If you can do this for 1/100th the cost of a proprietary solution, kudos to you. The lack of real support aside, the software pieces simply don't exist in OSS/GPL land.

Unclarity

[Eudial]

The unclarity!

These are just some of the questions popping into my head:
What is SAN?
What does it do?
How is it disruptive?
Who does it disrupt?
What does it store?

Can't say skimming through TFA makes it a lot clearer either.

Also, two obscure articles is media buzz?

Re:Unclarity

[Yetihehe]

Not every article is adressed to people which are not in field of storage...

Re:Unclarity

[TubeSteak]

What is SAN?
What does it do?
How is it disruptive?
Who does it disrupt?
What does it store? http://en.wikipedia.org/wiki/Storage_area_network
It's remote storage.
Their new tech saves you the trouble of swapping HDs.
It disrupts the people offering maintanence contracts.
It stores whatever you want.

http://www.xiotech.com/images/Reliability-Pyramid.gif
My question:
What is "Failing only one surface"

Re:Unclarity

[ILuvRamen]

the less they tell you, the more hype it gets regardless of how good it is. Remember Vista? It was supposed to be the end all OS sent straight down from heaven but they didn't release any specifics. And now look what happened. I doubt a magical storage system that can heal itself and never fails or loses data is a bit of an exaggeration too.

Re:Unclarity

[sadgoblin]

They should atleast give poiners at where to look... and I mean exact links, not "google it".

Re:Unclarity

[Xaedalus]

Assuming you're being literal with your confusion... A SAN is a Storage Area Network that organizations use to back-up up data off their main networks. A lay explanation: think of your normal network and how it's connected. A SAN network (usually composed of fibre channel or SCSI connections) underlays that existing standard network and moves all the data you want to back up to disk or tape, without eating up the bandwidth you have on your normal network. It's usually driven by a back-up server, or sometimes by a normal server (but only if you want to eat up your processing power). What this disrupts (if it's true) is how a SAN network monitors itself. It's basically pro-active monitoring and a different configuration of spinning disks. I'm not sure which RAID array they're using, so it may not be as 'revolutionary' as they're proclaiming it to be. Please note: any network admins, PLEASE feel free to correct me if I'm wrong (because there's nothing worse than giving a layman explanation that's inaccurate).

Re:Unclarity

[Anonymous Coward]

Just to clarify, SANs generally aren't used primarily for backups - they're just used for server storage to have a centralized and more thoroughly redundant setup than local disks (i.e. put your 40TB document repository on the SAN and connect it over fiber to the server, or have your server boot off the SAN instead of local disks, etc). They're treated like local disks by the OS, but are in a different location connected via fiber or nowadays via iscsi.

While you can sometimes do some neat tricks with backups and a good SAN infrastructure, it's by no means its primary purpose in life.

Re:Unclarity

[Anonymous Coward]

Let's see now... ah! I've got it. Here's an exact link for you: http://www.google.com/search?hl=en&q=san

Re:Unclarity

[DJProtoss]

You could use it for that, but thats not the main use.
It *is* a network like your ethernet network (with switches, adaptors, etc), but usually its a FC (fibre channel) rather than ethernet. You use a SAN to put your servers disks in another box to the server.
But why would I do that? heat, consolidation, redundancy.
A typical setup is to have a few 1u or 2u (rack heights are measured in u, which iirc is about 2") servers attached to a 3u storage controller.
This is a box with lots (typically 14 in a 3u box) of drives. There will be a small computer controller in there too as well as some raid chips.
Typically in a 14 drive box you might configure it as a pair of 5+1 raid 5 arrays and a couple of hot spares (5+1 means 5 drives of data and one parity drive). Effectively your 6 drives appear as one with 5x the capacity of 1 of the component drives. You can survive the loss of one drive without losing data. If you do have a drive go offline, the controller should transparantly start rebuilding the failed disk on one of the hot spares (and presumably raise a note via email or snmp that it needs a new disk).
The controller is then configured to present these arrays (called volumes in storage speak) to specific servers (called hosts).
The host will see each array as a single drive (/dev/sdX) that it uses as per normal, oblivious to the fact that its in a different box.
Now to revisit the why we do this:
1. heat - by putting all the hot bits (drives) together we can concentrate where the cooling goes
2. reliability - any server using the above setup can have a disk fail and it simply won't notice. With the hot spare setup, you can potentially lose several drives between maintainance (as long as they don't happen at once).
3. cost - you can buy bigger drives, then partition your array into smaller volumes (just like you partition your desktop machine's drive) and give different chunks to different hosts, reducing per GB cost (which when you are potentially talking about tera and peta bytes worth of disk space is rather important).
as for what these guys are up to, I've not had a chance to look yet. I might post back.

Re:Unclarity

[DJProtoss]

Ok, Have now rtfa'd. Basically, they have built a shiny controller/enclosure (the enclosure is the frame that contains the drives and the controller the circuit that interfaces, although to be confusing controllers often are built into enclosures (especially on the lower end) and still referred to as a controller)
This controller is a sealed unit (read: better heat/vibration support, but not a user servicable component) with excess disks inside (multiple hot-spares, so even if several drives fail over time it keeps going), combined with the knowledge san techs across the globe know: most errors are transient, and if you pull the disk out and stick it back in, it will probably work again. They have just built a controller that does that for you automatically. Definately on the evolution rather than revolution side of things, and I have to admit I fail to see the disruption here, although I could well be missing something ( the whitepaper is somewhat light on details shall we say ).

Re:Unclarity

[wgaryhas]

First thing I thought of was that when 1 platter in the hard drive fails, it still uses the other platters. But I may be wrong.

Re:Unclarity

[gnick]

It stores whatever you want. I'd like mine to store beer and bacon. Any idea on the capacity or replication capabilities?

Re:Unclarity

[Yetihehe]

I have seen some commentary on wired, which essentially sayed using internet and wikipedia can be compared to using brain enchancers, i.e. wiki user is expanding their brain with knowledge which he can "remember" just by typing words (like searching your memory for something you don't remember easily). This said, it looks like you are not using your intelligence and even don't want to augment it in any way. It would take you only 2 seconds to make a search, but you've chosen to remain in ignorance.

(* This is not intended as flamebait, use google people!)

Re:Unclarity

[Prof.Phreak]

indeed. servers that ``fix themselves'' with magic pixie dust... haven't we heard this 10 years ago from IBM?

Re:Unclarity

[Jesus_666]

Perhaps it somehow prevents a scenario where more than one platter is inaccessible at the same time? Although I guess you'd have to switch to single-platter hard drives in order to even attempt that...

Re:Unclarity

[DeadDecoy]

I'm sure this computer would do the job: Fridge Computer

Re:Unclarity

[silas_moeckel]

You forgot the most important bit speed. SAN's are orders of magnitude faster than most internal hardware raid. Think many GB's of battery backed up write cache.

Re:Unclarity

[jwgoerlich]

better heat/vibration support, but not a user servicable component

Heat is key here. Have you ever stood next to a petabyte of storage? Or even a few terabytes? Most Sans kick off a lot of heat from all those disks. When looking San to Hvac, 1 TB to 1 ton is typical.

Xiotech's ISE mounts the disks on a very large aluminum alloy heat sink. The heat is wicked away from the drives. This makes for better heat dissipation and less heat on the disks, thus improving cooling and extending lifespan.

Xiotech had a petabyte of storage on the SNW expo floor. I stood right next to it, surrounded by the crowd. The heat? Next to none. There was no additional cooling required for the demo either. It was completely ambient temperature. The cost savings in HVAC must be rather impressive.

J Wolfgang Goerlich

Re:Unclarity

[ErMaC]

Disk drives are made up of surfaces on platters. Generally, disk drives have multiple platters, with each platter having a top and bottom surface.

Currently, in a disk if one chunk of a surface has a problem, the whole disk is bad. The disk has no way to communicate which part has died.
Xiotech's hooks into the firmware allow it to write around bad areas on the surface of a disk, and when a portion of a surface does fail it only has to rebuild that portion, rather than the entire disk drive.

Re:Unclarity

[Anonymous Coward]

The largest configuration right now is 10 TB.

Re:Unclarity

[ILuvRamen]

lol I worked for IBM for 3 days through a contractor and they couldn't even pull some magic pixie dust out of their ass to get their hand scanner PDAs to work hehehe.

Re:Unclarity

[Chrisje]

You're so right.

As a Mass Storage man for HP, I was hoping to see something about cool Storage Area Network technology. Self-healing and fixing SAN's would indeed be a cool thing, because the way I see it I get more questions on the Infrastructure than the actual boxes in the SAN.

If you look at current offerings from a couple of the major vendors, you'll see that there are boxes that are already guaranteeing 100% uptime and have all the redundancies and diagnostics built in to actually deliver the goods.

The problem with Data storage hypes such as this (if it ever becomes a hype) is that people all of a sudden think that having such a box is a substitute for a decent support contract on the actual infrastructure or a backup.

Your DBA inserts incorrect data into a production database. Your exchange server becomes virus ridden due to insufficient patch management. Your users delete that project folder that is very important. Someone snags the cable in between the server and the storage device because they're having a bad day. Windows machines on a SAN do tape polling thus disrupting the bus.

The examples of configuration mishaps, logical data loss and sheer accidents far outstrip the instances where such a box would go down itself. So while it's nice that someone claims to have come up with a new Disk Box that will heal itself, the summary is misleading because it claims it will heal the SAN.

Which it won't, judging by the Marketing BS I just read.

Re:Unclarity

[Zashi]

You can also share drives between computers/OSes. This works out especially well for blade centers where you can share a base system install with 20 blades. Also, if a system goes down it's pretty easy to simply take out your Fibre Channel HBA and stick it in a similar or identical system and run just as you were before. It's too bad the 8gb/s cards cost ~$1,500.

Oh, on a related note, you can do FC over Ethernet. You can also do Ethernet over Fibre Channel. Yep. You can have 8gb/s TCP/IP connections. Or you can settle for a $10 2gb/s card. (annoyingly though, you can't find a fibre channel switch no matter what the speed for less than a few thousand dollars).

Re:Unclarity

[severoon]

Reading a site like this is going to expand your knowledge, but not by handing you all the information on a silver platter. All it can do is show you the door, but you are the one that must go through it.

Re:Unclarity

[I.M.O.G.]

HVAC cost would not be affected impressively the way you imply - you said you didn't feel the heat coming off, but that does not mean less thermal energy is being dissipated. It just means that same amount of thermal energy is being dissipated in a way thats less noticeable when you stand next to it. That energy will build up in an enclosed room unless proper HVAC systems are countering the effect.

Disruptive?

[99BottlesOfBeerInMyF]

The result is quite impressive, a SAN that can practically heal itself, as well as prevent common failures.

Maybe I'm missing something. I read their announcement and one of the articles on this new product. As near as I can tell they're selling SAN systems where instead of plugging in individual drives, you plug in a box with two drives in it. They paired this with some nice software for working around failed sectors and rewriting correctable drive problems. I guess I'm just not all that impressed. Is this really "disruptive" technology? It looks like evolutionary improvements and some nice automation to take some of the grunt work out of managing SAN.

I'm, admittedly, not an expert on network storage. So what do people think? Is this really the best thing since sliced bread or just another slashvertisement someone hyped to sound like news for nerds and rehashing a lot of marketing weasel words?

Re:Disruptive?

[cdrudge]

Actually, I think instead of a a box with one drive...or two...you will have 10 3.5" or 20 2.5". So you have one big RAID-like cluster with a big "gas gauge" like dial on the front that tells you how much performance you have left...whatever that means. Whoopdedo. But I think that since they use an acronym every other word in the ESJ article that we should be very impressed.

Re:Disruptive?

[kaiser423]

well, RTFA. For mod points, it's disruptive because it runs Linux!

The second article describes this very well. One big extra is that this system can perform all of the standard drive-repair operations that typically only OEMs can. This helps to keep you from replacing drives that aren't bad, but had a hiccup.

It's also not just two drives in an ISE, but more like 10-20 (3.5" and 2.5" respectively) with a bunch of Linux software to give each ISE a pretty robust feature-set in itself. Then they also up the block size to 520 bytes, leaving space for data validity checks in order to keep the silent corruption problem from sneaking into the system.

In the end, it's probably not wholly revolutionary. It does seem like an evolutionary jump though; with great performance, great feature set, and a very well thought out system that brings new technology and ideas to bear.

Re:Disruptive?

[timeOday]

So you have one big RAID-like cluster with a big "gas gauge" like dial on the front that tells you how much performance you have left...whatever that means. Whoopdedo.

I would call that a great thing. I've never understood why I couldn't just have a bank of a dozen drives with another 10 empty slots, and have it move data around automatically to increase performance and maintain redundancy. When enough data is stored or enough drives break that I'm close to losing redundancy, a light turns on, and I pop in another few drives and it keeps chugging.

Re:Disruptive?

[Anonymous Coward]

EMC Clariion storage arrays have been doing the 520-byte block thing for at least 5 years, probably longer. Perhaps they aren't using the ANSI standard, but it didn't exist when they started using the extra 8 bytes for checksums. It may go all the way back to the Data General days...

Re:Disruptive?

[Anonymous Coward]

NetApp and IBM/Engenio also have this sort of capability already too. And more; NetApp can even fix most cases where the disk reports success on a write back to the controller, but actually failed because of bad media issues (yes, disks can lie to you!). The only new thing here is the sealed unit, and XioTech will drown in their own RMA costs if their reliability is even just the slightest bit below projections.

Re:Disruptive?

[mochan_s]

I would call that a great thing. I've never understood why I couldn't just have a bank of a dozen drives with another 10 empty slots, and have it move data around automatically to increase performance and maintain redundancy. When enough data is stored or enough drives break that I'm close to losing redundancy, a light turns on, and I pop in another few drives and it keeps chugging.

One reason I can think of is because there is a high correlation of drive failures to the power supply and equipment that it's on. I've seen centers that have 1 rack unit where the disks keep failing.

Re:Disruptive?

[gfogus]

I have used both SAN and NAS. Here is the quick and dirty difference:

SAN - requires third party software to run. It gives slightly more options to the user, but it is not very interoperable. SAN software must be supported by the operating system.

NAS - is highly interoperable. Just type the UNC name and presto, you are viewing files. You can map the entire drive or select folders.

People have been making error correcting SAN boxes for ages, mostly in the form of RAID arrays. I wouldn't get another SAN box even if they were giving them away. Ok, if they were coming with hard drives inside, I would take them and rip the drives out.

If you don't have time to fool around with weird software, get NAS. It works.

Re:Disruptive?

[swb]

I've never understood why they don't do this. I went and looked at a Xiotech Magnitude in 2002 at their offices here in the Twin Cities. They gave me the big dog and pony show (my current boss had bought two a year before at a different company) and when they were demoing the unit, I asked if you put new drives in if it restriped the existing data to include the new drives to make adding new LUNs more flexible. They looked sheepish and said no, the new drives had to be created as a new drive group.

The SANs I've seen since then (admittedly all fairly low end, never gotten to use/manage one of the high and systems) all just look like direct-attach SCSI RAID with an integrated controller and a NIC/FC connector.

You would think the idea would be to chuck in drives (with some minimum, like 8 or 12) and have the physical data storage be totally abstracted from the user, with N+2 redundancy and hotspare functionality totally guaranteed, and then allow the user to create LUNs without concern for underlying physical storage.

When you need more space, you add more drives and the system manages the striping, re-striping as necessary for optimum throughput and maximum redundancy, rebuilding failed drives as necessary.

Re:Disruptive?

[igjeff]

You would think the idea would be to chuck in drives (with some minimum, like 8 or 12) and have the physical data storage be totally abstracted from the user, with N+2 redundancy and hotspare functionality totally guaranteed, and then allow the user to create LUNs without concern for underlying physical storage.

When you need more space, you add more drives and the system manages the striping, re-striping as necessary for optimum throughput and maximum redundancy, rebuilding failed drives as necessary. There are systems out there that do this sort of thing, but they're *expensive*.

Take a look at HP's EVA line. They're really quite good at this.

I'd be careful about using the terms "optimum" and "maximum" in that last paragraph, but they get quite close to that mark.

Other vendors have equipment that performs about as well...IMO, the HP EVA line is the best at it, however.

Jeff (only affiliated with HP as a mostly happy customer)

Re:Disruptive?

[johnlcallaway]

Having everything abstracted isn't always a good idea. Smart people will always be better at being able to get the most performance out of a machine, because we are creative and can think of how to use things in ways they weren't intended.

It's nice having the ability to abstract things completely when performance isn't paramount, but when those performance bottlenecks start to become an issue, it's nice to remove the abstraction and start becoming more specific about how things interact.

As a for instance, I remember a few years ago installing a SAN of sorts. The vendor's raid-10 wasn't fast enough, we would overrun the cache on bulk loads. So we built two raid-5s on two different controllers, and then 'mirrored' the controllers using the OS. Worked much better.

Re:Disruptive?

[swb]

Sure, nothing is as good as having a real human make the decision, but that scales really poorly.

Re:Disruptive?

[Mr2001]

Drobo does this. There's an LED gauge showing how full the array is, and when a drive needs replacing, the light next to it turns red. You can mix and match different sized drives, too - when it gets full, you can pop out a 250 GB and put in a 1 TB.

The downside is it only has 4 bays and connects via USB 2.

Re:Disruptive?

[ryguy]

I was involved with the evaluation, purchase and implementation of a Xiotech Magnitude 3D (a generation ahead of the Magnitude) a little over a year ago and I can tell you that they are able to add new drives to an existing disk group without a problem. You would be best off re-striping the drive group since you would have a group of partially filled drives in the group and another group of completely empty drives.

Their interface was quite slick and the technology was quite impressive. We compared them to IBM, Netapp, and Dell/Emc and we were much more impressed with the features and performance of the Xiotech. One of the features that most of the other manufactures could not do is striping across every drive in one array. (IIRC something like 200+ or so 15k SCSI drives could be striped in one drive group to produce incredible speeds)Using some performance testing tools (with only about 40 drives) we were able to see some speeds of up to 40x that of enterprise class directly attached storage on the same server.

The customer that they brag about the most is Microsoft who apparently has a couple of petabytes (at the time) of Xiotech storage.

Re:Disruptive?

[cthulhu11]

Back in the day when Fujitsu Eagles (anyone remember SMD?) were typical storage devices, it was entirely reasonable to map around a grown defect or two and keep running the same drive. With more recent disks, though, I've seen that factory error mapping is more reliable, and new defects on a disk presage a more widespread failure, so it's best to replace the whole disk before it crumbles. Given the way that modern disks have variable sectors/track and are addressed with logical blocks instead of by physical C/H/S, I don't see how it could be possible to avoid using a given head/surface. Rather than muck with all this added complexity, I'd rather just slap a bunch o disks into some enclosure or something like a Sun x4500, mirror them with SVM, and spend a fraction of the couple hundred grand saved on replacing my 1.6GHz G5 desktop.

Re:Disruptive?

[sincewhen]

That's a lot like the Drobo (http://www.drobo.com/) but it only holds 4 drives and connects via USB. A step in the right direction though.

Re:Disruptive?

[Phishcast]

Somebody already replied and said that what you're suggesting doesn't scale (they're right), but to add to that, here's why. When people invest in SANs and storage arrays, they typically want to take advantage of economies of scale and maximize utilization. That means they aren't going to put one application on the box and tune for it as you suggested, but they'll be putting 5, 20, or 100 applications on the same array. It's nearly impossible to tune each individual application's storage without dedicating spindles to each (or groups of each). Once you've done that, you may as well be buying separate storage for each application, and you're back to direct-attached storage without the economies and utilization you were after in the first place.

Striping all data across as many spindles as possible ensures reasonably predictable performance, and almost always better performance than you would have had by hand-tuning.

Re:Disruptive?

[Guspaz]

Unfortunately, it's extremely expensive and all reviews point to abysmal performance. Also, doesn't even have an ethernet jack to act as NAS. Sort of kills the deal, don't you think?

If, for that price (or less, those things are really overpriced), that had onboard GigE and fixed their performance problems, it'd probably be a decent deal.

Re:Disruptive?

[Anonymous Coward]

Not 5 years, since the early 90s! Welcome to the 20th century Xiotech!

Re:Disruptive?

[Mr2001]

Well, right, it's not really meant to compete with NAS. I think the performance problems are caused by the USB link. There is an attachment you can use to connect it to a network, but it's also expensive and doesn't improve performance.

But if someone could convince them to put their technology to use in a professional storage array instead of a consumer-level RAID For Dummies...

Sweet...

[steppin_razor_LA]

The disk healing features are very interesting.

We have a Xiotech Magnitude that we paid ~$150K for in 2003 that is sitting around like a giant paper weight. Any takers? $3,000? $2,000? going once... going twice... :)

Re:Sweet...

[schklerg]

I've got 3 collecting dust! And based on my experience with that SAN, I will never entertain the slightest sales pitch from any Xiotech rep. I'm sure they've gotten better, but rebooting after changing the contact info in the system is a bit absurd. Not to mention that the management / configuration was on a single IDE hard drive running MS-DOS. Since a reboot cleared all logs, tech support's stock answer for odd issues was, 'it was in a bad state'. Had it moved to Arkansas? BAH!

Re:Sweet...

[Anonymous Coward]

Ive got an old Magnitude that needs a forklift to remove, the monster had a microcode flipout in 2004 and corrupted 2 TB of data and Xiotech told us it was a server problem....really Windows, Mac, and Linux servers all simultaniously failing, I think not. Stay as far away from Xiotech as you possibly can!!

Re:Sweet...

[steppin_razor_LA]

'it was in a bad state'. Had it moved to Arkansas? BAH!

HAHAHA :) :)

My biggest bitch was that the $150K solution consisted of $37K of HW and the rest was software licenses and/or support that doesn't seem to be transferable. This basically means that there is no secondary market for the devices because anyone who would buy one would need to buy new software licenses. Since the SW licenses are more valuable than the HW, it wouldn't make sense to buy used HW. "nice"....

The above weighed in heavily in our decision not to go with Xiotech for our second SAN.

That said, the article was still interesting. :)

Re:Sweet...

[Jesus_666]

Well, you could rip out most of the internals and put generic PC components in. Then you end up with a 37 Kilodollar case mod.

Re:Sweet...

[Moekandu]

So, if you don't mind my asking, who did you go with instead of Xiotech?

Re:Sweet...

[schklerg]

I think you were asking steppin, but as another ex-Xiotech user I thought I'd add to my earlier post. We went to a midrange IBM platform, the DS4500. For low to moderate workload it's been pretty good. But code update are disruptive and it's really made by Engenio so real problems (which we've seen) end up getting redirected to them. Some of the issues (and our own growth) have lead us to now replacing the DS4500. Upgrading SAN's isn't fun, but now we're looking at the enterprise offerings from EMC, IBM, HDS so I'm hoping this next SAN will be more stable and perform better so it can stay in place for more than the 3 years it sits on the books.

Move along, nothing to see here

[Spazmania]

They've integrated the controller and drive into devices that consume 3U of space in a rackmount computer cabinet. So now you can't upgrade a drive, you can only replace a module. Brilliant.

The only thing this is likely to disrupt is Xiotech's cashflow.

Re:Move along, nothing to see here

[Chas]

If (big IF), the units eat drives (as in kill them in a non-repairable way), yeah, it's a bold and stupid move.

If (another big IF) the unit keeps soft-failed drives (which weren't really bad to begin with) in play longer because it can recover them from *burps* in the system, then it's entirely possible that the unit could potentially be a money-saver.

Re:Move along, nothing to see here

[ErMaC]

You don't need to replace a module, because it doesn't break. See the failure rates/service event numbers from their presentations.

People are so used to disks failing. Disks shouldn't fail as often as they do, and most of the time they don't fail at all - the storage controller is at fault because the drive and the controller have such a limited language (SCSI) to talk to each other with. ISEs do away with this limitation.

Re:Move along, nothing to see here

[MadMorf]

You don't need to replace a module, because it doesn't break. See the failure rates/service event numbers from their presentations.

Lab failure rates mean very little.

Didn't Google just blow the lid off of the disk manufacturers MTBF numbers, by reporting their own failure rates as being an order of magnitude higher?

Wait till they have a few thousands of these deployed. They we'll kknow how good they really are...
And, that's when companies with big money to spend will take notice.

Re:Move along, nothing to see here

[Spazmania]

The odds of the soft fail savings catching up to the difference in economy of scale are not good.

Re:Move along, nothing to see here

[tppublic]

Honestly, there isn't much cash flow to disrupt. This isn't EMC, HP, IBM or Hitachi.

The purpose of this product isn't to penetrate large data centers... of if Xiotech thinks it is, then they need new marketing employees (and quickly). Large data centers HAVE the expertise on site to do individual disk replacements, and those large enterprise data centers will demand the feature sets that exist in the much larger equipment from the larger vendors named above.

This is targeted at much smaller data centers, probably those with very simple SANs (think a dozen or two servers), where the data center management skills won't match those in the larger data centers (simply because you have one or two generalists, not a dozen+ specialists). For those smaller sites, the return on investment for a system that requires less maintenance (and also less expertise) may make sense...

Yes, this is evolutionary from a technical perspective, but it still approaches the solution in an interesting way... and may find its own market niche.

Re:Move along, nothing to see here

[Spazmania]

Thing is, I spent the last couple years playing this game. I started with a dozen 36-gig scsi disks that had bad sectors on them. I did thorough tests abandoned the whole gigabyte where the bad sectors were found and software-raid-5'd partitions from multiple drives, skipping those bad parts.

Guess what? It didn't work out. The bad zones spread and they spread faster than the the raid software could detect the new failure and rebuild onto the spare.

I quite enjoyed the experiment, but these were on my home servers. I wouldn't dream of doing this in a production environment. When the raid controller kicks the drive for -any- reason, it's back to the manufacturer for warranty replacement. The data is far to valuable to play games with it.

Re:Move along, nothing to see here

[ToasterMonkey]

Big storage vendors are expected to charge less for maintenance now? Where, What, Huh?

Seriously, it's not like an enterprise disk array owner can just stop over at Best Buy, pick up a new drive and pop it in whenever he feels like it.

Sure the price of DISKS will go down but you know the cost of having some monkey stop by the data center to replace a failed DMX drive isn't going anywhere.

Supposedly the maintenance from Xiotech is going to be $1 on these things. Gimmick, sure, but in theory that's where the savings are supposed to come from.

Re:Move along, nothing to see here

[Guspaz]

You'd detect the failure the instant you tried to read the sector, and could speed up this process by using idle time to check sectors for corruption (checksums help). With double parity, you'd have to have the same stripe go bad on three different disks to lose data, and I have trouble believing that the failures would spread that fast.

I mean, I still wouldn't trust it in a production environment, as you said, I just wonder how a more dynamic system (RAID-Z, for example) would handle the same scenario. At the very least if you decided to simply try to get one clean copy of all data off the disks to migrate it to backups to prevent data loss.

As in, your test would be constantly reading all data on the disk in a loop, and seeing how long it would take before data was lost permanently.

Looking forward to my Chihiro drive!

[RobertB-DC]

So if San is the technology, the drive that implements it would be called Chihiro, right?

Oh, that was Sen . My bad, sorry.

(Well, it makes as much sense as anything. It's not like I'm going to bother reading TFA when it's clearly marked "hype".)

Obvious press release

[Anonymous Coward]

Too bad the marketers who wrote it don't know how to speak English. Otherwise, we'd be able to understand and be excited about this.

Re:Obvious press release

[Ungrounded Lightning]

Too bad the marketers who wrote it don't know how to speak English. Otherwise, we'd be able to understand and be excited about this.

They speak great management buzzword. But not tech - buzzword or otherwise.

astroturfing

[nguy]

I don't see anything particularly "disruptive" about this. Lots of storage systems are "self healing" and based on hot-swappable elements.

The whole thing sounds like astroturfing.

Re:astroturfing

[soontekoh34]

I don't see anything particularly "disruptive" about this. Lots of storage systems are "self healing" and based on hot-swappable elements. The whole thing sounds like astroturfing. How is hot-swapping a drive self healing?

Re:astroturfing

[K. S. Kyosuke]

How is hot-swapping a drive self healing? The engineer swapping the disks is chained to the unit, has no name, sleeps in a doghouse and was a part of the delivery.

Re:astroturfing

[nguy]

How is hot-swapping a drive self healing?

It isn't. Who said it was?

The company claims two innovations: self healing and rapid replacement. Both are commonplace.

Clear now?

Re:astroturfing

[soontekoh34]

This is what you said, "I don't see anything particularly "disruptive" about this. Lots of storage systems are "self healing" and based on hot-swappable elements." I don't know of any self healing storage systems let alone "lots" of them. Can you name some?

Re:astroturfing

[Anonymous Coward]

RAID systems are "self healing: when a disk goes bad, the system keeps going with the ones it has, and when you replace the bad disk, it fills in the missing bits. If you run RAID on a SAN, RAID systems can automatically swap out bad disks for good ones (simply by switching to a spare drive).

At a lower level, there is also a lot of "self-healing" technology in drives: bad block and bad track replacement, sophisticated error correction and recovery, automatic drive adjustment, drive health monitoring, etc.

The notion that "self healing storage" is something new is ridiculous.

Tired of overused buzzwords

[93 Escort Wagon]

Why do people keep referring to incremental improvements to existing technology as "disruptive"? It's pretty obvious people don't understand the phrase "disruptive technology".

My favorite misuse was when a marketing droid referred to Intel moving from a .65nm fab to .45nm as "disruptive". It's not just marketing folks, however - I've heard engineers and even my own college professors (usually if they're trying to turn their research into something commercially advantageous) do this.

Re:Tired of overused buzzwords

[melted]

>> referred to Intel moving from a .65nm fab to .45nm as "disruptive"

Disrupted AMD pretty good, from what I can see.

Re:Tired of overused buzzwords

[Idiomatick]

Its pretty obvious you don't understand how resistant markets are to change. Anything new and not garbage that escapes past the board of directors is shocking and disruptive.
 
Sorry about the pessimism(its exam week)

Re:Tired of overused buzzwords

[networkBoy]

it's 65 and 45nM or .065 and .045uM
pedantic, I know...
but .45nM would be phenomenally disruptive as it would literally be two orders of magnitude better litho than what is currently attainable commercially.

-nB

Re:Tired of overused buzzwords

[rcamans]

At least it disrupted 10% of AMD's workforce (They are getting laid off)

Compellent

[Anonymous Coward]

I'm not sure why this announcement is really news...somewhat interesting though, is that many of the founders and former employees of Xiotech have left to start a company called Compellent http://www.compellent.com/. Compellent's disk technology, imo, is a lot slicker than Xiotech's, particularly their "automated tiered storage".

Re:Compellent

[Anonymous Coward]

Compellant is a dead duck. You can trade bubblegum for their shares at the moment.

Re:Compellent

[medelliadegray]

I admin a compellent san, and i find their "automated tiered storage" to be great in concept, marketed superbly, yet highly expensive and highly lacking in configurability.

If you want data automatically moved down to a slower tier, but it gets touched just once a day. Good luck in getting it to move down automatically.

I anxiously await the day when the SAN market is acknowledged as the scam it is (a glorified raid controller), and the various SAN companies die off in droves or become an everyday appliance they really are. It's obscene paying a grand for a run of the mill sata disk, and additionally paying about as much or more than the disk in various licenses. All the while gouging you yearly for 'support' contracts which are a sizable fraction of the cost of both hardware/disks/and licenses.

Hurray for hemorrhaging cash!

Re:Compellent

[Anonymous Coward]

OK, now you're knocking bubblegum.

Re:Compellent

[steppin_razor_LA]

The idea that decisions can about what disk/RAID class is best can be made by the HW on a block by block basis is very slick.

We didn't shell out the $s for the licenses because the old model (i.e. my databases are RAID-10, my file servers are RAID-5, etc) works "good enough" when compared to the sticker of the automated tiered storage licenses.

Re:Compellent

[jwgoerlich]

If you want data automatically moved down to a slower tier, but it gets touched just once a day.

Data progression does the moving. DP only runs once a day by default, but you can change this schedule. You can also kick DP off manually. How? Ask Co-pilot.

J Wolfgang Goerlich

Re:Compellent

[Anonymous Coward]

Did you look at the specs and data sheets on xiotech's site before the diarrhea of the mouth kicked in? It is quite a radical concept, re-manufacturing drives in place, allowing the failure of single platters while keeping data on the other ones!

Re:Compellent

[Anonymous Coward]

You should take a gander at mainframe software. The govt and big companies pay _buttloads_ of money for glorified secure ftp, tape management software, the ability to use a tape for more than one job, job schedulers, and a lot of other stuff we really take for granted in the "open systems" world.

I think this goes for all IT products: the price generally goes up exponentially from home, small business to enterprise, while the quality per dollar is pretty much a flat line with a nose-dive at the end. Sure, more money gets you bigger stuff, but it aint anything like the auto industry I guess I'm trying to say.

Re:Compellent

[Anonymous Coward]

http://datacore.com/products/prod_SANmelody.asp

Worked for us. we had similar issues about costs with our HP SAN.

Much cheaper than a proprietary solution and we can use whatever disk we like. ISCSI or FC SAN, doesn't matter. Support is pretty good too, and I believe (I don't sign the checks) that the costs are very reasonable compared to a 'hardware' solution.

Re:Compellent

[Lxy]

Considering that Compellent and Xiotech both run Seagate disks, how is the "disk technology" better in Compellent? Keep in mind that Compellent is merely a consumer of Seagate disks, Xiotech now owns a piece of Seagate that makes this new SAN do what it does.

Also, doesn't Xioetch do automated tiers of storage (ILM or some weird acronym)?

Btw, I don't work for either of these companies, but I have evaluated both products extensively.

Re:Compellent

[Lxy]

I wouldn't call a SAN a "scam". You get what you pay for.

Want a big box full of disk, fully redundant (RAID 1,5,5+1,10, etc)? What it cheap? Got spare parts? Then, my friend, FreeNAS is for you. A homebrewed SAN that delivers enterprise capable performance for practically nothing.

Oh, you want FAST disk? OK, then you have to shell out for SAS or FC disk. Can still use FreeNAS, but now your hardware costs have gone up.

Your box has multile SAS controllers, multiple SAS drives, and now what? Gotta go external. More parts, more cost.

Want professional support on that FreeNAS box? Ummm... whooops. OK, then instead of FreeNAS let's go with Lefthand networks. Similar product to FreeNAS, but carries a price tag for support.

Wait, we just spent a lot of money to build a SAN to our liking. And it's still just a bunch of crap I cobbled together. May as well just have gone to a vendor like Xiotech, Compellent, NetAPP, Equalogic, whoever and bought the real thing.

Then enter larger corporations that have demand for high performance, full management, automatic everything, and of course a 24x7 tech-lives-onsite kind of contract. That's where EMC, HP, and the like have their products.

So you see, there's an option for every need. There's an option for every budget. Some people have a legitimate need to drop some cash on a big reliable SAN. Some don't. Pick your price, but don't call it a "scam".

Hypocritical Reluctance

[Anonymous Coward]

On the Xiotech site:

"#1 Lowest cost per disk IOPS
#1 Lowest cost per MB/sec"

Looking around, I don't see any quoted prices on the page.

It's funny how it's always a project in itself to find the price tag for products. When companies run on "the bottom line" why are they so reluctant to tell us what the consumer's "bottom line" is straight forward and upfront?

It should become law; that to advertise a product, you must post clearly what the price tag (range) is either at the top or bottom. Especially if you are telling people its "cost-effective" without providing the cost. Am I saving $1 or $10,000?

Captcha: increase

Re:Hypocritical Reluctance

[Anonymous Coward]

The storage industry is notorious for trying to hide their price lists. Check out http://storagemojo.com/storagemojos-pricing-guide/ for street prices on storage gear. It's not all up to date, but you can get a ballpark without requesting all sorts of quotes from a reseller.

Re:Hypocritical Reluctance

[sarabob]

There's a link in the white paper to the benchmarks (http://www.storageperformance.org/results/benchmark_results_spc1#a00064), which then gives you pricing info on the tested configurations.

A 1TB array with 40, 15k 2.5" drives in raid 1 is $36,500 (list price is $61k, the price used by the spc has a 40% discount!) with a three-year, 24/7 4hr maintenance contract. It generates 8,720.12 SPC-1 IOPS, making it $4.19/IOP

The other tested config used 20 146GB drives to get 5,800 IOPS for $21k, $3.53/IOP.

(a 12TB netapp system, FAS3040 gets 31k IOPS for $420k = $13.61/IOP as a comparison, no 40% discount here :-)).

Now double check the quotes. "World Record SPC Benchmark 1T: Lowest cost per SPC-1 IOPS1". Hmmm. RamSan400 gets 291,208 IOPS for $194k, at $0.67/IOP (Some places on the xiotech website say 'lowest cost per disk IOPS' as some kind of get out clause, but not all.)

Interesting that the support contract for netapp & EMC appears to be as much as a xiotech array + contract, although they do seem to have 140-150 disks in the tested configurations rather than a measly 20 :-)

Re:Hypocritical Reluctance

[Anonymous Coward]

People who believe SPC benchmarks are going to really show you an even-handed apples-to-apples comparison of storage arrays are fools. It's a vendor tool, they basically get to design their own tests and configure their boxes to optimize the results. If you dig deeper into each test you'll see boxes configured in ways you'd never actually have them in production.

It's akin to NVidia or ATI doctoring drivers to excel at specific benchmarks.

Re:Hypocritical Reluctance

[Anonymous Coward]

The RamSan400 has a maximum of 128GB in it for that price also. The Xiotech device used 20 146GB disks (3TB) in it so the capacity right off the bat needs to be considered at the same time.

Re:Hypocritical Reluctance

[flying_solo]

People who believe SPC benchmarks are going to really show you an even-handed apples-to-apples comparison of storage arrays are fools. It's a vendor tool, they basically get to design their own tests and configure their boxes to optimize the results. If you dig deeper into each test you'll see boxes configured in ways you'd never actually have them in production.

It's akin to NVidia or ATI doctoring drivers to excel at specific benchmarks.

While this is often true (often configured for tests in ways never used in production), if you actually read the whitepaper, it clearly states that the Xiotech test had the disk 75% full. Many other SPC tests "cook the books" by using only "outer tracks" on the disks, running the disk only 10% or 25% full to reduce seeking or keep all I/Os in cache etc. But when you run the disks 75% full you are essentially full-stroking the disks, so are demonstrating much closer to "real world" use, aren't you?

Re:Hypocritical Reluctance

[flying_solo]

The last statement "Interesting that the support contract . . ." - where can I find more information on that? I thought Xiotech had announce a 5-year warranty, with no hardware maintenance contract of any kind required on this?

You MUST be new to storage technology.

[crovira]

I'm not going to educate you except to tell you to Google for it.

The disruptive part is that it seems to be much more reliable which would mean that you can wave the tech goodbye for a while, instead of having to lose access to a sting of drives RAIDed together while they have to rebuild a drive which failed and needed replacement.

Think of running XFS without having to worry about the drives' physical reliability because they're really reliable. (If you've got 5PB online it usually "which drive just failed", instead of "here's the data")

"What does it store?" Jeez ... "Shoes" What the hell do you think it stores? How about data!

But you are correct in that TFA didn't carry a price list of various configurations.

Re:You MUST be new to storage technology.

[Phishcast]

I'm not going to educate you except to tell you to Google for it. The disruptive part is that it seems to be much more reliable which would mean that you can wave the tech goodbye for a while, instead of having to lose access to a sting of drives RAIDed together while they have to rebuild a drive which failed and needed replacement.

Umm...I think you forgot what the "R" in RAID stands for. You may have somewhat degraded performance during a rebuild when you spare in for a drive which has failed, but you don't lose access to any data because of a single disk failure (Save for RAID 0, which isn't really RAID to begin with).

I wouldn't call this disruptive. It sounds like they've done some smart things to bring disks back to life when other hardware would call them failed, but you can bet that they're packaging more spares in these non-user serviceable enclosures than you would in a user serviceable configuration.

Re:You MUST be new to storage technology.

[afidel]

They have 20% spare capacity which IS a lot more than you would put in a user servicable enclosure. The great thing is you get the speed advantage of the spares because while they limit the amount of data use to 80% they are normally using all spindles in the pack.

Evolutionary, but on multiple fronts.

[khasim]

They've added the OEM services to THE DEVICE ITSELF. (evolution)

They've made those OEM services on the device AUTOMATICALLY kick in. (evolution step 2)

They've sealed the units. (evolution)

Which, in effect, means that most of the SAN expertise that FORMERLY required an experienced tech is now incorporated and these SAN's can be installed and "maintained" by less technically skilled personnel.

Which will make these devices VERY easy to sell. You pay ONCE for the tech and save on the cost of the technician's salary.

I will be watching for these in the future. IF they are as good as they claim, I will be buying three of them.

Re:Evolutionary, but on multiple fronts.

[Amouth]

I already requested a price sheet - the sales rep I talked to didn't know what it was and is going to give me a call back..

we shall see if they are affordable

Re:Evolutionary, but on multiple fronts.

[Anonymous Coward]

But wait, before you buy that. I am a storage admin and I can just put some cheap drives in a RAID and watch them for you. You will save so much money on hardware this way.... ;)

guess it's time to go back to school huh......

Please don't disrupt my storage

[Anonymous Coward]

Xiotech Unveils Disruptive Storage Technology Oh noes! Klingons with disruptors are in the server farm! Scrape them off Kirk!

"a SAN that can practically heal itself"

[SwashbucklingCowboy]

Well, no. It's an array that can practically heal itself (at least in theory). BIG difference...

Not disruptive, in fact, dated technology

[mollog]

12 years ago I was on a team that developed a RAID system that managed itself in this same way. I don't know why Slashdot would bother to post an article about a disk system and not provide any sort of details about what is new. Automatic rebuilding? Active spares? Management of a heterogeneous set of drives (mixing drive capacities, for example)? In fact, when I left that company seven years ago, they had on-the-fly volume 'snapshot' capabilities for backup, journaling, and other uses. Built in diagnostics were included. Where's the beef?

Re:Not disruptive, in fact, dated technology

[gweihir]

Quite true. LLinux software RAID also offers all this stuff today. Thus story is nothing byt badly researched hype.

Re:Not disruptive, in fact, dated technology

[Knara]

100 or so engineers involved in the project have replicated Seagate's own processes for drive telemetry monitoring and error detection -- and drive re-manufacturing -- in firmware on the Linux-based ISE. ISE automatically performs preventive and remedial processes. It can reset disks, power cycle disks, implement head-sparing operations, recalibrate and optimize servos and heads, perform reformats on operating drives, and rewrite entire media surfaces if needed. Everything that Seagate would do if you returned a drive for service. My software RAID definitely doesn't do that.

Re:Not disruptive, in fact, dated technology

[gweihir]

Well, not the low-level stuff, which may not be a good idea anyways, since drives change and what about replaing a disk in there in a few years? The righ strategy is to drop defective drives and replace them from spares. Trying a few resets is fine, IMO, and the kernel does this anyways. If a power-cycle is needed, I would classify the disk as defect, same with the other close-to-hardware stuff.

As to prevention, Debain, e.g., runs a RAID check every 30 days by default. I add a long SMART selftest every 14 days or so as cron-job (or anacron job for machines not always on). You can have these more often, if wanted. Personally I would expect that the stuff ISE does in addition will not recover most defective drives anyways.

Wrong metaphor

[bperkins]

Each datapac in the ISE presents a gas-gauge-like monitor showing you how much performance is being used [...]

If you can use your gas gauge to measure how fast you're going, you're probably driving too fast.

Nothing to see here, move along...

[MadMorf]

Nothing really new here, except the box is sealed, which means when they have bought a batch of disks with an undiscovered flaw, there's no way to fix or replace them...

Seagate Tomcats anyone?

Also, would you trust your enterprise storage to laptop drives? Running 24/7/36...

How long will those last?

Hell, most SATA disks are unsuitable for anything but nearline storage, and even then, they're iffy...Keep plenty of spares!

Re:Nothing to see here, move along...

[Phishcast]

Well, saying laptop drives isn't really fair. They've been making 10,000 RPM serial attached SCSI disks in a 2.5" form factor for quite some time, I know Sun uses them in servers. I'm not sure if there are 15,000 RPM disks out yet in this size or not. These are not your 5400RPM laptop drives.

Re:Nothing to see here, move along...

[MadMorf]

Well, saying laptop drives isn't really fair. They've been making 10,000 RPM serial attached SCSI disks in a 2.5" form factor for quite some time, I know Sun uses them in servers. I'm not sure if there are 15,000 RPM disks out yet in this size or not. These are not your 5400RPM laptop drives. Ok, fair enough.

Still, I personally don't like "black box" systems...

Re:Nothing to see here, move along...

[Torg]

I tend to agree. Take for the example of "Self Healing". It sounds nice, but it is not the SAN it heals its just the array. And more often the SAN itself has problems then the arrays. This would be a mis-configured SAN, with single attached hosts and hosts without multipathing software. Or an array mis-attached to the SAN.

Disk drives fail, and fail more often in batches by lot then by mistaken identity. When your drive tells you it is going bad it is so that you can get the data off safely, before it dies. Not mark one platter bad and use the rest. So do you really want your disks to lie about problems?

But no, wait, we will add some flashy lights and guauges. They are about as useful as the oil light on your dash. Without context and without knowing overall trends they are useless.

Use a few technical terms, well mis-use them. To this add falsely what is the bane of SAN admins (no it is not "No Fault Found"), and you have a marketing release about an array.

Make an array that can tell me "look idiot you plugged both HBAs into the same switch". Oh wait, that is called SRM software.

Re:Nothing to see here, move along...

[cthulhu11]

Indeed, the 2.5" SAS enterprise-class disks that Sun ships eg. in an x4100 are pretty sweet.

Re:Nothing to see here, move along...

[afidel]

You can definitly get the SAS drives in 15K, I use them for my Citrix blade servers (HP BL460C). During heavy login periods the 15K is a must, especially since the blade only allows two spindles.

Re:Nothing to see here, move along...

[neurovish]

Also, would you trust your enterprise storage to laptop drives? Running 24/7/36...

How long will those last? I have a lot of HP bl30p servers setup with software raid1 arrays on their internal disks...which really are laptop drives. I'm amazed that we still haven't had one of them fail these past 3 years.

Of course, now that I say this, 20 of them are going bad tomorrow.

Just marketing redefining words.

[Ungrounded Lightning]

It's pretty obvious people don't understand the phrase "disruptive technology".

Hey!

  - If Microsoft's marketing department can redefine "Wizard" from "Human computer expert acknowledged as exceptionally skilled by his peers" to "only moderately brain-damaged menu-driven installation/configuration tool",

  - why can't Xiotech's marketing department redefine "disruptive technology" from "quantum leap in price/performance ratio of a competing technology based on a massively different architecture that makes it out-compete and displace the previous market-dominating solution" to "incremental generational upgrade in the latest model of our product which we hope will convince you to replace the competitor's product with ours (and disrupt both his business plan and your IT operation)"?

Re:Move along, nothing to see here, but shame

[Anonymous Coward]

Indeed all the ideas put forward by Xiotech are suspect or old
Re-manufacture a drive with problems - not a good idea. Let Seagate provide you a replacement and if it really didn't fail then why did you call it out as a failure?
Old - spares in place, controllers by the drives, data integrity checks, etc all done before (and probably better).
Only thing worse is to buy a system from Compelent.

You'll notice they have No OEM sales - only end customers. OEM's have the time to test a product.

Other things along these lines

[TheSync]

In the broadcast engineering space, we see a lot of this kind of thing...

Avid Unity ISIS
Omneon MediaGrid
DataDirect S2A

Re:Other things along these lines

[csoto]

I am tempted to mod you down just for MENTIONING the damn Unity thing! :b

Re:Other things along these lines

[TheSync]

I am tempted to mod you down just for MENTIONING the damn Unity thing! :b

Hah, I'd totally deserve it!

This exists, and has for a while

[Phishcast]

Off the top of my head, all of the following companies have storage arrays which basically do exactly what you're asking for. When you create a LUN it's across all available spindles and data will re-balance across all available disks as you add more, all with RAID redundancy. I'm not sure about N+2 at this point, but RAID-6 is becoming ubiquitous in the storage industry.

HP (EVA)
3Par
Dell/Equallogic
Compellent
Pillar
HDS (USP)

I'd be shocked if Xiotech doesn't do this today.

Re:This exists, and has for a while

[swb]

The Equalogic we have at work doesn't work like this, or at least it doesn't seem that way to me. LUNs don't seem striped across all disks from what I can tell, but I haven't worked with it in a while.

Re:This exists, and has for a while

[Phishcast]

Well, I specifically asked this question of an Equallogic salesman at the bar last year and he told me this was how it worked. He may have been blowing smoke or not understood the question. Either way, he was buying :).

Re:This exists, and has for a while

[swb]

He had the right answer to "Are you buying?" at least.

Re:This exists, and has for a while

[Anonymous Coward]

Compellent definitley does this already and has for quite some time.

Course you can do all that with a LAN too.

[Colin Smith]

I'm seeing less and less need for multiple networking technologies. IP uber alles etc etc.

Failing only one surface (was: Unclarity)

[jwgoerlich]

What is "Failing only one surface"

A hard drive can fail in many ways: sector, track, platter, head. ISE can fail just the one surface -- say, a platter -- and keep writing to the remaining device. The broken platter is removed from service while the remaining disk storage continues to be used until end of life.

This is all done automatically and transparently. What they are trying to eliminate is the time it takes for someone to physically swap out a disk.

J Wolfgang Goerlich

Why wouldn't I just use RAIDZ ?

[Anonymous Coward]

Why wouldn't I just use boxes with a bunch of SATA drives and run OpenSolaris with RAIDZ and two hot spares? I would save money and even if I have so many boxes that I have to replace disks every day, it is still feasible. The huge amount of money saved will more than cover the cost of having an admin swap in new disks.

And if you're talking about a small SAN, then RAIDZ is a no-brainer.

ZFS (with RAIDZ) is also available for Linux, OS/X and FreeBSD, but for production use, OpenSolaris and FreeBSD are the only two choices.

Re:Why wouldn't I just use RAIDZ ?

[ePhil_One]

Why wouldn't I just use boxes with a bunch of SATA drives and run OpenSolaris with RAIDZ and two hot spares?

Because of the overhead of designing such a system? I'm a pretty storage savy guy and I would'nt have a clue where to start home building the system you describe. Until there's a well designed "Live CD" type install that nets a simple to use appliance type interface, this is not a viable alternative for most shops.

That said, this this seems liek a nightmare. all the drives are sealed into "Drive Pacs" and can't be replaced individually? They are putting a lot of faith in their rebuild capability...

Re:Why wouldn't I just use RAIDZ ?

[ckaminski]

www.openfiler.org
www.freenas.org

HTH. HAND.

Re:Why wouldn't I just use RAIDZ ?

[saleenS281]

http://www.nexenta.com/corp/

Re:Why wouldn't I just use RAIDZ ?

[ePhil_One]

What the parent was posting about was the potentially very powerful combination of ZFS and OpenSolaris, which would bring all sorts of well established, stable, volume management goodness. I'm aware of both these products, and have considered using them, but while they deliver similar end products, they are distinctly NOT the RAIDZ solution the grandparent was going on about.

That said, the next reply seemingly IS, except for the 1TB limitation of the "free" version.

But thanks for pointing those out, I've actually been eyeing them for a while as a way to address some storage challenges I'm facing. I remember being disapointed by some implementation details, but I no longer recall what they were. Do you have experience w/ either product?

Re:Why wouldn't I just use RAIDZ ?

[Anonymous Coward]

ZFS on FreeBSD isn't a wise choice for a production system yet. Try it and you'll see. It's got a ways to go yet.

Re:Why wouldn't I just use RAIDZ ?

[ckaminski]

RAIDZ/ZFS... I like them, I like that this functionality is incorporated into Solaris in such a clean fashion. But it's no different than LVM/MD in Linux. I would like to see the Openfiler admin bits stripped out of it's LiveCD/rpath environment and adapted for Solaris - I think there's value to be had there.

I've played around a bit with OpenFiler in a VM with the NAS and it was about what I expected for a web interface over Samba and NFS. I'm a bit more interested in the iSCSI functionality to serve as my base for a VMware/Xen/Solaris lab I manage, which I haven't tried yet.

Re:Why wouldn't I just use RAIDZ ?

[ePhil_One]

But it's no different than LVM/MD in Linux.

I've worked with Veritas's VM, IBM's VM, Linux's VM, and Linux's MD, I'm quite confident they are NOT the same. Perhaps they all provide redundancy, the the LVM's all slice and dice the disks in a similar manner (Multi-Disk is a fundamentally different product), but the Linux versions had a fair bit of catchup to do to get close to the commercial products in terms of both usability and functionality.

I've got 7 years experience managing SAN products from EMC & IBM, though I have to admit after playing w/ Equal Logic's web interface I'm in love. The Solaris/ZFS combo sounds amazingly powerful & stable, and several Slashdoters have pushed it as "why buy a solution", but the learning curve seems orders of magnitude steeper than Linux with far more serious repercussions for screwing it up, meaning its only really viable w/ a user freindly interface and simple install info. I'd really love for thsi to be a viable option, if nothing else so I can get I cheap iSCSI system up for development & test environments (I have the CPU power, but not the disk I/O I need)

Re:Why wouldn't I just use RAIDZ ?

[afidel]

Can your homebrew solution scale from 1TB to 1.something Exabytes with 64GB of cache all without any downtime? Oh and does it come with a 5 year warranty? Also what do you do when your box is at a remote datacenter on another continent and you can't get ahold of an operator?

NAS is big again

[heroine]

This is so arcane. It's like sitting on Tasman Dr. watching Net Appliance buy up lot after lot & VA Linux just announced a new thing called a build to order NAS. NAS is the future again. Buy Excite.com!

Saw This Unveiled at SNW in Orlando Today

[Anonymous Coward]

They made a large splash at SNW today in Orlando. They had a petabyte of worth of these running. I am not convinced. Storage is so competitive and vendors are constantly touting some gain over another. The storage appears to be RAID 10 and yes all the drives are in two sealed packs inside the 3U enclosure. Interesting, but not disruptive.

Missing something...

[Anonymous Coward]

The real news isn't the "box of disks with controller" part, there are plenty of others doing that. It's the other stuff they've thrown in. Other vendors only do some of these. DataDirect, for instance, can only power cycle a whole drive after trying multiple times to read the data. Xiotech's stuff can fail an individual drive head if it needs to. There's also nothing to suggest DataDirect takes the disk's geometry into account in their RAID implementation, like Xiotech does. Basically, the smallest unit is just a whole disk to them (and every other storage vendor on the planet), so their best effort is to cycle the whole drive. I'm not even going to touch the other two vendors you listed, those are cheap junk. Seriously, a 1u "storage" blade with two disks and ethernet connectivity? Dude, Avid Unity is reselling cheap blade servers with a clustering filesystem (which is probably some off the shelf stuff from a certain Linux vendor).

Now add on-board disk remanufacturing. You read this correctly: drive remanufacturing on the device itself. Steve Sicola (he ran ASA and is now CTO at Xiotech), Richard and Ellen Lary (Ellen is now a vice-president at Xiotech), and the other 100 or so engineers involved in the project have replicated Seagate's own processes for drive telemetry monitoring and error detection -- and drive re-manufacturing -- in firmware on the Linux-based ISE. ISE automatically performs preventive and remedial processes. It can reset disks, power cycle disks, implement head-sparing operations, recalibrate and optimize servos and heads, perform reformats on operating drives, and rewrite entire media surfaces if needed. Everything that Seagate would do if you returned a drive for service. The headline reads "...Disruptive Storage Technology" because of how incredibly low maintenance these supposedly are. The story goes that Seagate was afraid to market it directly because certain big array vendors threatened to stop buying their disks, so they sold the IP to Xiotech.

I know I sound like a shill so I'm posting AC. You need to really watch out with these "no-name" storage vendors. Even the big name ones are full of shit, so the little ones are almost pure snake oil. I know, Xiotech is pretty unheard off too, you have every right to be weary. Look how excited they are to drop a name like Seagate, hehe.

Monty Pedantic

[epine]

it's 65 and 45nM or .065 and .045uM
pedantic, I know... Not nearly pedantic enough, unless you knew Herr Meter personally. Tell me, Herr Meter, why were you named that way? What did you discover to make yourself famous? And why did you say that Ångström deserved what he got?

Funny, I was thinking on the way home that the intergalactic subway machine in Contact blew up because the alien schematic contained a typo calling for 1 eV, and the people building it failed to read it as 1 exavolt.

Wikipedia tells me that 1 EeV/c = 1.783×1018 kg Really? I thought 1 EeV would be more impressive. I guess it's not the exotic extraterrestrial vroom vroom I thought it was. No, wait, what am I talking about, vroom vroom is v.

Re:Monty Pedantic

[epine]

Whoops, Slashcode consumed the second dimension and more. That's probably where the alien error originated in the first place. Morse coded the schematic onto a six sided cube without first pressing submit. Cocky bastards.

Maybe this will help you nerds understand

[Anonymous Coward]

Watch this video then you will understand why this is a revolution. Storage that doesn't fail.
http://youtube.com/watch?v=Rj_3S9-2g4Q&feature=PlayList&p=F2903C9E7C88C543&index=0