A Look At the Fastest IDE Drive Yet
muks writes "Here's an article on Tom's Hardware about IBM's Deskstar 75GXP. It has some good points on why we still need UltraATA/66 and faster IDE interfaces while hard drive transfer rates don't keep up. "
← Back to Stories (view on slashdot.org)
Yes, this is an older design (Tom reviewed it in May), and it's based on the 440BX chipset. Asus apparently got around the "no ATA/66 on 440BX" by using a different chipset for the controller -- and put in two of them for good measure!! Cool.
I haven't looked around much, but with the proliferation of IDE/ATAPI Zip drives, DVD drives, CD-RW drives, and the incredible availability of cheap hard drives, I have to think this idea [of multiple IDE controllers] is going to stick around. Other motherboard manufacturers should get the clue very soon, if they haven't already.
"...America's great minds of today, teaching America's great minds of tomorrow. Poor bastards." -- A Beautiful Min
Dare I suggest that the software was your problem? I've never understood how CD-Rs are supported in WinNT. It seems like you install one, and windows creates a bunch of drive letters for it. Then if you actually want to burn, your burning software has to grab the device some other way. Ugh.
In Linux, I've never had the first problem with disks, scanners, cdrom, cd-rw, and the rest competing. I've had devices that just weren't supported properly, but that never had the effect of destabilizing the entre bus.
Frightening, no? I remember thinking I was ripping off the vendor when I bought a 420 MB IDE drive for $300. Under a buck per meg, I was in heaven...
--
The unsig!
Your facts are a little crooked there. A RAID of five 9 GB drives could have a capacity ranging from 9 GB (bizarre 5-way redundancy), to 45 GB (no redunancy, but very fast). My point still stands: you can used SCSI equipment with the same or higher capacity and vastly superior performance, for the same price as a new ATA drive.
What about drives? Drives report geometry in 3D anyway. You'd have to change the firmware of all drives, or implement similar kludges in all new harddrives. In the end, it is probably too much work for too little gain.
A deep unwavering belief is a sure sign you're missing something...
Not really. If you haven't got enough RAM, a hard-drive twice as fast wouldn't improve performance. What really improves, however, is general usage. For example, when I upgraded to my Matrox 20giger (the previous fastest harddrive before the IBM) from my old 6.4 GB one, Quake didn't run any faster, but stuff like recompiling X or the kernel, unzipping tar files, starting up, etc, all became much faster.
A deep unwavering belief is a sure sign you're missing something...
we didn't have these fancy namby pampy hard drives. We had cassete tape recorders to load and save programs.. And we liked it..
Now I feel old.
I used to be a SCSI bigot but I could not justify the difference in price for desktop machines.
A typical example:
Quantum Fireball lct10 20.4GB EIDE, price $99
Quantum Atlas III 18.2GB Ultra Wide SCSI hard disk, price $269
Before someone pulls out the respective spec sheets for each drive and starts quoting MTBF numbers and data transfer rates, let's look at the big picture here:
The average user - heck, even the power user, is not going to see much of a difference between these drives in day-to-day use. Yet the SCSI drive is > 2.5x the price of the EIDE drive.
SCSI is a mature technology. Even EIDE drives use the SCSI command set over the EIDE bus. So WHY do we still see these huge price differences?
I hate to nitpick, but if there's no redundancy, it's more of an AID than a RAID, isn't it? :)
"Destroy science and religion. Science would re-emerge exactly the same; but not religion." - Penn Jillette, paraphrased
Look, SCSI does this too. To the BIOS, the linier SCSI drive is mapped into the 3D geometry, and mapped back into linier by the OS. In fact, I think IDE drives these days do that too. Also, there is nothing wrong with IDE. Sure it is popular to curse IDE, but if you do, you're not using it right. With the advent of UltraDMA, IDE drives have as low a CPU usage as SCSI drives, and with UDMA100, IDE drives have more bandwidth than all but Ultra160SCSI and FibreChannel. IDE drives still have lower TPS, but on a desktop machine, that really doesn't matter. This is proven by the new IDE drives (like the 75GXP.) The 75GXP actually performs as fast as some 10,000RPM SCSI drives on the outer tracks. IDE really isn't a bottle neck in this situation.
A deep unwavering belief is a sure sign you're missing something...
UDMA66 can have 4 devices per port. So, 8 devices on a standard system. Now, if they would just put more ports on a board, so overhead per chain stays fairly low, that'd be nice.
Or even a more wild dream: Make CPUs more efficent, instead of just faster.. Naa, never happen.. =]
75 Gig hard drives! (I feel really old, sometimes.)
"..don't you eat that yellow snow."
Complaints about an obsolete, bottlenecking technology on /.? Increadible? When somone says "X should DIE!" People say "what for, it works!" When people say "IDE should DIE!" people say "YEAH!"
PS> X should DIE!
A deep unwavering belief is a sure sign you're missing something...
SCSI, hands down!
My home machine is all SCSI, all the time. I'm using a Mylex/BusLogic BT-958 controller (does Ultra 2 Wide). On the internal bus, I've got two Seacrate 4G drives, an IBM 8G drive, and a Plextor 40X CD-ROM. These are connected together using a 68-pin twisted-pair cable. On the external bus, I have a Castlewood ORB drive (if you've got one of these, download the v3.0 SCSI firmware now!), connected using a 68-pin shielded cable. I used to get occasional bus lockups, but once I replaced the Toshiba CD-ROM drive with the Plextor, they went away.
My laptop (HP Omnibook 800CT) has built-in NCR/Symbios SCSI controller. It rocks.
The IDE controller in my home desktop machine is unused, and will remain that way. IDE sucks. Always has.
The reason I insist on SCSI is mostly because I've been a SCSI wonk since my Amiga days. Sticking with SCSI lets me use all my old drives (and get the data off them!). All the old drives still work, including a really massive 5.25" full-height 600M drive from HP that needs a power supply all to itself just to spin up. Being able to get at all these drives, both on the desktop machine and the laptop, is insanely useful.
It wasn't so long ago that SCSI and IDE drive prices were at near parity. The SCSI drive had maybe a $10 premium over its IDE counterpart. But now you can expect to pay up to 50% more for a SCSI drive. This is clearly opportunism on the manufacturers' part, as the electronics aren't sufficiently different to warrant such a price disparity. They're simply soaking the server market.
Grr, I say, grr...
Schwab
Editor, A1-AAA AmeriCaptions
You answer your own question. I don't agree with you, but for the simple fact that Joe-Bob-Computer-User isn't going to 'see much difference' in the drives is the exact reason. Well, okay, actually isn't because they don't know the difference so they won't pay more, therefor prices aren't going down.
.. The simple point of the reduction in overhead, is more then enough for me to spend a few extra dollars (Well, maybe not a few) in getting SCSI. Now, I just need to get those few extra dollars.. =]
Rememeber the very *very* slim profit space on HDs, and then ask a company who does SCSI how much they make off them. I'm sure it can't be too much. Think Flat-Screen TV/monitors here...
On the other hand, I don't agree with you one bit that Joe-Bob-Computer-User won't notice a difference. Let's dream for a second here. Joe-Bob buys a nice new Compaq with a expensive SCSI setup in it. You looked at a pre-fab computer lately? They load so much crap at Windows startup it's not funny. It's pure evil, I saw a Athlon 1ghz with a GeForce run slower then my K6-2 500 with a TNT2.
So let's say that startup crap is gone, and the OS is properly tuned. If you don't see a difference, then you have ADD.. =] The difference is very clear, try burning a CD on a nice SCSI controller with a nice drive. Then try loading some programs while doing it. Not much difference then normal use. Now try doing that on a IDE drive. Oh dear, you have whacked your IDE bus, and the overhead is causing your computer to run like a TRS-80
Then again, this is just my observation.. =]
SCSI drives report the total number of blocks (READ CAPACITY command). It is the SCSI BIOS on a PC that translates that into cylinders, heads and sectors.
Mea navis aericumbens anguillis abundat
>No, but there aren't any devices out there that are going to suck up all that bandwidth on the consumer side of the market, either.
A disk -> disk copy (for example - large zips and encryption (lots of temp space)m compiles, mp3 moves) are all quite a bit faster
>I dare you to try and get 15 SCSI devices from different manufacturers and different scsi revisisions running reliably on the same bus.
Well, here's what I have connected presently:
IBM 9GB 10krpm - SCSI-3 UW (almost 2 yrs old)
Seagate 4GB 7200rpm - SCSI-3 UW (4 yrs old)
IBM 18GB 7200rpm - SCSI-3 UW (2 yrs old)
Plextor PX-40TW CD - SCSI-2 UW (1 yr old)
Plextor PX-W8220T CDR/RW - SCSI-2 UN (1 yr old)
Nec 6x CD - SCSI-2 N - (real old)
Iomega Jaz 1GB - SCSI-2 N (4 yrs)
Iomega Zip 100MB - SCSI-2 N (3 yrs)
Running on an Adaptec 2940UW (I don't remember the BIOS revision)
(my Wang DAT drive isn't currently hooked up in that system)
Never a SCSI problem... never... not in NT, not in Linux, not in BSD...
>It'd be great if devices would disconnect.
Umm.... they do.
>Do you know how hard it is to find scsi cables and adapters in most cities - and how much you'll pay for them? UGH.
Never had a problem... and it you want a cable that has 15 connectors and an active terminator, you are going to pay a little more for it.
Oh, yeah - I left out external connections and chaining in my earlier list...
I've never had any problems with my SCSI connections (Adaptec, Tekram, Future Logic controllers on a PC, IBM controllers on RS/6K, some sort of strange controller on a Sun...), except when the RAID controller on a RS/6K went belly-up after 5 years, but they installed a new one that same afternoon, and all was well.
--
"It's tough to be bilingual when you get hit in the head."
I didn't say windows, I said "the software", which includes the vendor's drivers.
I agree completely. A cheezy hack that's held on far too long.
Unfortunately, really big scsi drives are more expensive than really big ide drives.
If i wanted 30 gigs for mp3 storage or something similarly trivial, i'm afraid it'd be IDE in a heartbeat.
keep in mind, my workstation at home has five (5) discrete scsi busses, and no ide drives. But I've got less than 20 gigs of scsi storage total.
At least, with storage arrays getting bigger and bigger, the old UW stuff is hitting the market pretty cheap occasionally.
This is just like television, only you can see much further.
I've got a yamaha burner as well, and it works great, but the one thing that I noticed a few years ago -- there are IDE burners out there that work just as well for less. I think they've got an advantage in the "value" pc, which seems to be what most "consumers" are buying...
And as far as adaptec cards? They are good, but I think they're slightly expensive... of course, the cables are what will get you, if the card doesn't. I've got a modest LAN at home, and while I've seen the virtues of SCSI, I have no need for it at home as far as a storage solution -- IDE with 2-3 users seems to work just fine.
Also, I bought a promise ata-100 card for 40-50 bucks about 2 months ago, and using the "one drive per controller" method seems to make linux a little more responsive.
Karnal
Look at a used computer shop or sale. I have gotten an EISA Adaptec SCSI-2 dual channel card that I haven't tested (among the computers I have I don't have an EISA MB), but it was before I bought it. Other SCSI-2 cards I got there have worked fine. They had 1Gig SCSI hds (as-is) but the 4 I got 3 worked) for $2.
I have gotten lots of $5 or so Adaptec SCSI cards, and $20 or so NCR ones, that work with no problems. The sad thing is they are going out of business, because a few things investers wanted didn't turn out right, not that they were losing money.
What sort of disconnect are you talking about? Are you talking about the protocol-level disconnect, which when the device does so, it lets the other devices in the system use the bus as well, or do you mean the "hot-swap" disconnect, which requires special enclosures?
Granted, both are very nice; my all-scsi box at home outperforms (in IO) the all-IDE box at werk; similar configurations on both boxen.
--
SCSI has its own problems. Expensive cables, termination hassles, broken firmware, a physical interface that seems to get redesigned every year.
Mea navis aericumbens anguillis abundat
1)Demand increases. 2)Prices and profit margins go up. 3)Everybody and his brother moves into that business. 4)Supply increases. 5)When the supply increase catches up to the demand increase, it often overshoots, leading to a glut. 6)Prices go down. Simple.
/.
/. If the government wants us to respect the law, it should set a better example.
Well, that guy was talking BIOSs and all, so I was assuming the discussion was about PCs.
A deep unwavering belief is a sure sign you're missing something...
Very nice... now how about some price comparisons:
IBM 20GB 75 GXP $136
IBM 60GB 75 GXP $417
Seagate 18.4GB X15 $482
(figures from pricewatch)
So that's over 3x the capacity for LESS THAN the
same price... SCSI rocks if money is no object,
but it'll still be a while before I build a PC
with it...
.technomancer
.technomancer
Yes, SCSI kicks the crap out of IDE, no argument there. But it is significantly more expensive. Like the other guy said, if you can afford it, great.
"Destroy science and religion. Science would re-emerge exactly the same; but not religion." - Penn Jillette, paraphrased
I was about to add a big cheap IDE drive to my PC, but a few smaller SCSI drives off e-bay are a much better idea!
Heh, my antique PII 266MHz only has a 4GB UW SCSI drive right now...
I'll save the usual rant about why SCSI is better then IDE. I've already posted it to Slashdot like five times since I joined. It's getting old. But I will respond to a few choice comments...
... doesn't work so well when the devices start to compete and see who's smartest ...
... how much you'll pay for them?
No, but there aren't any devices out there that are going to suck up all that bandwidth on the consumer side of the market, either.
Why is it that all the IDE apologists out there say "Home users don't need that kind of bandwidth!" in response to us SCSI users, when the topic that spawned the thread is invariably a new, faster IDE spec?
If home users don't need the speed, why do they keep making IDE faster?
I dare you to try and get 15 SCSI devices from different manufacturers and different scsi revisisions running reliably on the same bus.
Well, I've never gone all the way to 15 different SCSI OEMs on one bus, but I have combined several, many times, without problems. I've also seen the occasional incompatibility due to buggy firmware. But then, I've also seen the occasional incompatibility with IDE, and with only two devices. Hmmmm.
It'd be great if devices would disconnect..
It would be great if they powered on, too. Good thing they do!
Um, do you have something in mind here, or are you just making it up as you go along? Seriously. The above makes almost no sense.
Doesn't happen with IDE.
Because it cannot. How can an inherently dumb bus have problems with intelligent devices?
Do you know how hard it is to find scsi cables and adapters in most cities...
Oh, for crying out loud. Why don't you just say you don't like the acronym or something? If you can't find SCSI cables, you're looking in the wrong places. I live in East Buttmunch, Cow Hamster (AKA New Hampshire, a rural state in the North-East corner of the USA), and I don't have any problems finding SCSI cables.
This is mainly a matter of economy of scale. If they made as many SCSI cables as they did IDE, SCSI would be about that cheap, too.
dragonhawk@iname.microsoft.com
I do not like Microsoft. Remove them from my email address.
How do you fix the problem and maintain backwards compatibility? The BIOS could be modified to look for new signatures in the disk and operating system boot blocks. If it found the new signatures, it would use the redesigned code. If not, it would fall back to the current kludge. You would need Intel and Microsoft's support to make this work.
Mea navis aericumbens anguillis abundat
www.enthea.org
Actually this isn't the case, as holographic means that means the data is stored everywhere but in increasingly lower resolutions as you chop it up. That means access times are almost instantaneuos. According to Imation, they have already been able to achieve throughput in excess of 60GB/sec. For obvious reasons though, they are keeping the specifics of how it works under wraps. They are confident that a commercial holographic storage product should be available by the end of 2002.
www.enthea.org
Why would anyone ever need...
...a 75GB hard drive? (2000)
...a 2GB hard drive? (1994)
...a 100MB hard drive? (1988)
... more than 640K of RAM? (1982?) ^_^
More processor speed, more RAM, more disk space... someone always opens their yap about how we won't need it, and then a year later they can't imagine what they did without it, but are still bitching about how uneccesary the _next_ upgrade is. There's always a use for it! If there doesn't seem to be, it's because we didn't have the resource to play around with before.
I swear. When cold fusion is perfected, these same peolpe will be running around going "Why would anyone need all that energy?!"
The enemies of Democracy are
I found a 3' Ultra2 LVD cable for sale here for only $45 dollars. 6' VHDCI SCSI cables are going for $93 at the same place. Still inifitely less than a 3' or 6' UltraATA cable.
For internal use, a 22-inch Ultra2 cable costs $17, which an 18-inch UltraATA costs $2. Neither price is likely to put off most computer buyers.
Finally, comparing current SCSI prices with current ATA prices is bogus, because the current SCSI product is many time better than the current ATA product. It's apples to oranges. If you find SCSI drives that perform on par with the fastest ATA drives, which is not easy, you will find their prices are reasonable. In fact, even crusty old SCSI drives like the Seagate Barracuda 4LP, introduced almost 5 years ago, still take the fastest ATA drives to town. Surely you could find a farm full of used Barracuda 4LPs for only a few hundred dollars, string them together into a software RAID array, and have vastly superior disk throughput than you would if you bought a shiny new IBM Deskstar GXP.
And haven't you ever mounted a drive (or an array) externally?
have more than 4 IDE devices without going to absurd lengths to do it. Nowadays, most folks have at least two CD type drives (a burner, a reader, a DVD, etc.), maybe an IDE tape, at least a couple of hard drives, and not enough money to go SCSI on everything. I don't need the arguments about SCSI; I accept that it is better, faster, more scalable, etc., but I don't accept the needless price gouging. I want cheap, relatively fast, and most importantly, more than 4 IDE devices.... Surely that's not so much to ask. If IDE is that bad off, let's quit using it
Windows is going the way of phlogiston...
We need to use a better bus/protocol. SCSI has a lot of advantages over IDE, and a couple of disadvantages, but here's my quick list:
Pros:
1. You don't waste all of that theoretical bandwidth like you do with IDE. A more streamlined protocol.
2. Despite the smaller protocol, it happens to be a lot more flexible - types of devices, NUMBER of devices, error recovery...
3. Did I mention that a UW SCSI bus can have 15 targets (i.e. drives) with one initiator (i.e. controller), while IDE is still stuck at 2 per bus...
4. Disconnect.
5. Smarter devices - they do more for you.
6. There is no pro #6.
7. Generally better warantees on hard drives (ok, not a technical detail, but a selling point).
8. Cable length.
9. CPU utilization (much better than ATA DMA)
10. It's fun to say.
11. Cheaper than fiber channel.
12. You can run it as a lan...
Cons:
1. More expensive controller.
2. More expensive drive electronics.
Still, well worth the money... plus, the better drives are always out for SCSI first, since they *are* the drives that go in even small servers (not your linux box/webserver/desktop/ftp server/Win98 box - a real server).
--
"It's tough to be bilingual when you get hit in the head."
On of the big sellers over here on the Mac end of things are IDE->Firewire cases. The flexibilty of Firewire devices and the cheap price of IDE drives.
More info is often posted over at Accelerate Your Mac web page.
Of course this brings the ire of us folks who want native (ie bridgeless) Firewire drives, but hey, you can't expect everything at once, right.
Pope
Freedom is Slavery! Ignorance is Strength! Monopolies offer Choice!
It doesn't mean much now, it's built for the future.
I realize that comparing used to new is unfair, but i'm not the one who brought up the price to performance ratio. The best bang for your buck is by far used SCSi drives.
I'm about ready to buy a new system and recently looked into moving from SCSI to 1394... forget it. I still want 1394 for it's multimedia potential -- yes, those SONY Digital handycams with 1394 look real promising. But fiberchannel, 1394 ain't.
Frankly, why don't people use a two tiered approach with their systems. Folks here always degenerate the argument into an either/or debate. Either Linux or Windows, either GNOME or KDE, either IDE or SCSI; what bullshit. I use SCSI for boot and swap, and IDE for cheap storage. Do I need gigs of mp3's on a fast boot disk? No. Do I want my OS and swap on a set of chained master/slave IDE drives, with all the known contention this involves? No.
Use both to your advantage. You'll love the system speed of SCSI for what really counts, and the savings of IDE for what's a little less important.
- Total I/O per second: X15 117% faster
- Total throughput: X15 116% faster
- Average I/O response: X15 53% faster
Thanks to Storage Review for the benchmark figures.And remember kids, maximum length on those UltraATA cables is 18 inches(!), versus a luxurious 12 meters for Ultra160 SCSI, which transfers data much more quickly anyway.
N GHz CPUs
N Gbyte/s busses
N Gbps networks
N ns memory
and millisecond access on disk drives...
Government of the people, by corporate executives, for corporate profits.
First of all disk cache isn't high speed cache like your CPU uses. Its cheap standard dram type memory.
You don't need 2mb cache. Its pointless on modern drives. What you need is at least two full virtual "tracks" worth. Since most OS's want to format the drive so that its less than 1024 tracks that means a large ide disk will need 2*63*255*512=~16m bytes just to hold a virutal track mapping. If the drive doesn't have that, then when you do a large request, it will span real physical tracks and it will take several revolutions to get your data while you wait.
Of course this much memory would add nearly $16 to the price of a modern drive.
SCSI was supposed to be a general purpose interface where you could connect Drives, printers, tape drives, scanners, mice, kb's.....just like USB is pitched now. Except that it just never worked that way. Anyone tried to connect a SCSI scanner to a loop with a bunch of drives or even mix and match devices on a single loop? Sure it MAY work but you'll probably spend some time swapping out adapters, adjusting the connection order (even though that isn't supposed to happen), terminating, not terminating blah blah blorgh.
.5 meter external U2W 68 pin cable I bought was ~$75.00 retail. Sure I could have bought it mail order for about ~$25-30 but I can't be sure they'd send me the right part.
And oh yeah - - I've got a carton of SCSI 1,2, FW, Differential, "3" drives and adapters that don't work with one another, terminate differently have different cable connections, use 8-16bit interposers that don't work and generally require cables that cost of forking FORTUNE; The last
SCSI is great if you've got the money and time to set it up and support it but for home use? Bleh.
The old (E)IDE/ATA vs. SCSI flamewar again. I can't believe it! All I have to say is BAH
Each have their own applications.
As far as new equipment goes, SCSI is FAR more expensive. Yes I know you can buy old 9GB SCSI drives for peanuts and hook them up in a RAID and beat the pants off of any IDE out there. What about new systems? What if you dont have space or cooling for 8 cheap scsi drives?
What about devices such as the TiVo that demand 30 gigs of storage in a single drive (not to mention the other electronics) to make a profit at $500? Try doing that, SCSI!
After this, it may sound like I'm pro-IDE but really I prefer SCSI. But I'm not a bigot either and like to give IDE its fair due.
I am quite impressed with ATA/100. I set up a RAID-0 with the HPT370+Raid ATA/100 controller which came built into my new Abit mobo and two Maxtor 20.5GB ATA/100 Drives with 4M cache each. There is one drive per chain per channel. Ive got 41GB of storage now and I'm getting burst transfers from cache of 170-180MB/s! (In other words, my max transfer is better than Ultra160 SCSI) Also, I will never need to add extra drives etc in this particular system, so SCSI doesnt have it on me there. To top it all off, all the files on this system that are getting shuffled around are in the neighborhood of 100K. It's all run through the cache. SCSI would have cost me an extra $1000 to get the same performance. That's all there is to it.
~GoRK
Ok, first off - I used to be a SCSI fanatic, my first hard drive was a 80 Meg ST1096N I bought with 2 years of savings for an Amiga 500. SCSI kicks ass in the server room, but it absolutely sucks for home use. Why, you might ask?
1. You don't waste all of that theoretical bandwidth like you do with IDE. A more streamlined protocol.
No, but there aren't any devices out there that are going to suck up all that bandwidth on the consumer side of the market, either.
3. Did I mention that a UW SCSI bus can have 15 targets (i.e. drives) with one initiator (i.e. controller), while IDE is still stuck at 2 per bus...
I dare you to try and get 15 SCSI devices from different manufacturers and different scsi revisisions running reliably on the same bus. I had a couple different hard drives and CDROMs on a wonder scsi bus - it was a wonder why and when it would work.
4. Disconnect.
It'd be great if devices would disconnect..
5. Smarter devices - they do more for you.
Great idea - works well with scanners - doesn't work so well when the devices start to compete and see who's smartest, which is usually done instead of them doing what they're supposed to. Doesn't happen with IDE.
8. Cable length.
Do you know how hard it is to find scsi cables and adapters in most cities - and how much you'll pay for them? UGH.
12. You can run it as a lan...
Now this is cool; We did this with some amigas back in the day. Of course, ethernet cards work better now :).
Overall, my experience with SCSI was hell on earth. Trying to make a scsi cdrom, burner, hard drive(s), and other goodies from different manufactuters work and boot Windows was hell on earth; I haven't had any such problems with IDE, the controllers are there, and I don't need more than 2 hard drives and 2 removable media devices in any box (I'll just buy another box and set up a server).
YMMV though.
...the problem isn't hard drive protocols. Even with SCSI you're still in the millisecond-range for disk access and latency. What we SERIOUSLY need is some hardcore development in permanent storage with micro- or nanosecond seek/latency times. Like that holographic storage we always keep hearing about... you'd think after 20 years of research, someone would have a sellable version out.
"Destroy science and religion. Science would re-emerge exactly the same; but not religion." - Penn Jillette, paraphrased