Slashdot Mirror
IDE, SCSI And Recording Everything
Raju writes: "For many years we were told that SCSI is superior to IDE. I always made my systems with SCSI and the others in the household got el-cheapo IDE disks. In the past SCSI beat IDE hands-down but now according to Simson Garfinkel, "today's IDE drives are significantly faster than SCSI drives". In the article at O'Reilly Network he talks about the tests they had run for storage of network data on disks. In the light of this article does anyone see any reason for going with SCSI in a desktop machine? For servers with heavy disk usage patterns it might be different due to command queuing." Disk types aren't what the article's really about, though -- it's a top-level look at network forensics (including advice on building a traffic-analysis system), and makes some interesting points about the unbalanced growth of storage and bandwidth.
I have two sets of IDE controllers on my system. Each disk I have has its own channel and controller. Because I get to use cheap IDE disks, the cost is much lower than SCSI and the performance is right on par with it. Its not the technology -- its how its applied and used in real life.
--Kevin
SCSI is a very reliable system and is capable of handling high load from multiple requests. Never had the original CD-R writting issues that IDE had because the system was designed to handle constant through put. But like all things with price, IDE's affordability has won out over technical achievement and has slowly worked its way up, if you buy the right hardware. This is no different than from USB. Originally, it was pale compared to firewire. But, cheap gave it the edge to become a dominant standard. With USB 2.0, Firewire will be religated to niche before dying quietly. After all, for the price of a SCSI drive, I can easily buy three IDE's. It makes more sense to just mirror the data. However, where performance is the last word, SCSI will stay for now. At my work place, the amount of volume of data we process is staggering. There just isn't IDE hardware that handle the number of disks we need to have online.
My two cents,
Bel, the mostly sane.. "Of course I can't see anything! I'm standing on the shoulders of idiots." -- Me
As for arrays, beware of the benefits of striping. RAID 0 (striping) has the problem that the more drives you add, the less reliable your array becomes. RAID 0+1 (or RAID 10) mirrors the data as well and keeps your data secure in the event of a single disk failure (and RAID 10 can occassionally suffer multiple disk failures).
For the past five years I've run my system exclusively with SCSI components. When I first went out and bought a SCSI controller and a disk I paid a fortune for the privilege. At the time, the UW controller cost me $150 and a 9Gb IBM drive ran me another $300. The controller's SCSI BIOS added another 5 seconds to my boot time, and the IBM drive was full-height and loud as hell on account of it spinning at 10,000rpm. Regardless, I was a happy camper. I had consistently fast disk access, low latency and--best of all--I didn't get those annoying entire-system pauses while waiting for disk accesses to complete!
Over the years the benefit running SCSI decreased. First bus-mastering IDE channels came along and got rid of the annoying pauses. Then they started turning up the clock speeds with UDMA 66, 100, and so forth, until my aging SCSI drives could barely compete with even an average IDE drive.
Naturally, I did what any self-respecting bithead would do: I upgraded my SCSI components. By that time (circa 1999) the price gap between IDE and SCSI had narrowed somewhat (this was before IDE storage prices bottomed out) and I was able to purchase two 18Gb SCSI drives for a mere 25% than the equivalent IDE drives would cost me. And once again, I was happy with decent performance, low latency and high throughput.
Two weeks ago, I found myself scrabbling to free up a few megs and realized it was that time again, time to upgrade my storage. Looking at Pricewatch, I noticed that IDE drives are now cheaper than Big Macs and come in similarly absurdly-sized portions. Would you like 160Gb of space for your MP3s? No problem--they've got you covered, at $200 a pop! Meanwhile, relatively few vendors have stayed on the SCSI bandwagon, demand for SCSI drives is mostly limited to legacy systems that don't support an IDE bus, and a 160Gb SCSI drive will cost you $900.
In the face of this incredible price ratio, I did what any self-respecting bithead would do: I threw in the damn towel. Now I'm in a transitional period where I run 36Gb of fast UW SCSI storage and 160Gb of even faster IDE storage; I have a SCSI DVD-ROM drive, a SCSI CD burner, and an IDE DVD+RW burner, I/O controllers are fighting each other to the death to secure an interrupt, and the inside of my case looks like the aftermath of a tragic explosion at a cabling factory. I'm damned lucky my system is water-cooled, because I doubt any system fan could pull enough air through that morass of ribbon cables to make a difference in cooling.
The moral of the story: SCSI had its glory days, but it just ain't cost-effective anymore. And with Serial ATA looming on the horizon and promising God's own transfer rates, it just doesn't make any sense to buy SCSI.
Is that they can have a lot more devices, and it isn't just limited to storage. MY personal system has 5 HD's, 1 cdrom, 1 zip, 1 scanner. All scsi. If it wasn't for that one thing, yea, ide would make scsi people stupid now days.
Not to mention, there is also the smartness of the scsi controller. I've used "good" usb scanners, and ittakes over you computer when scanner. With scsi, you can burn a cd, scan a picture, and play quake3 with out a hickup. Now, who would do all that? Well think enterprises. Think 14 15000 RPM scsi drives in a raid 5 (or what ever). Or think media people having to render imamges while saving to a file and other stuff.
Oh ya, and nothing like sending in an older (3 year old ) scsi drive for RMA, with no questions asks other than "how can I help you".
But I have to admit, these days I keep quesioning myself on my coninuation of buying scsi for home. The I just look at all the things I have in scsi, and think of how I would "try" to do it w/o it. And I can't.
Oh ya, somethign that you ide people can't do
1 10k rpm hd OS
1 10k rpm hd swap/tmp
1 10k rpm hd data
1 10k rpm hd applications/games
1 10k rpm hd mp3/downloads, etc
The spirit of resistance to government is so valuable on certain occasions that I wish it to be always kept alive
Although many people discuss the superiority of the SCSI protocol vs. the IDE protocol, this is not really the question.
Manufacturers produce the fastest disks on the planet on SCSI interfaces only. There are no 10K/15K RPM IDE discs, period. If one wants the lowest access time available today coupled with respectable transfer rates, one must purchase a 15K RPM drive, which are only available in SCSI interfaces.
For single-user access patterns, the author is correct to state that IDE drives have the lead today. StorageReview.com recently reviewed the latest 7200 RPM Seagate SCSI offering, and it was beaten down in single user tests by half a dozen of the newer IDE drives; however, when tested with server access patterns, it was the clear leader (excluding higher-RPM offerings, of course.) Still, 7200 RPM drives can't beat 15K RPM drives in any access pattern.
And I noticed the author was RAIDing drives -- 3ware's RAID products are very high quality, and their performance exceeds each and every other RAID card out there, SCSI or IDE interface. That surely contributed to his conclusion that current IDE drives are faster than their SCSI counterparts.
Join the NFSNET. Our prime goal is making little numbers out of big ones. http://www.nfsnet.org/
On the surface, I would agree with you. However, the planned usage of the disk space in question becomes an important point.
I had this conversation with Greg Oster, a friend from University, who wrote the NetBSD RAIDframe implementation. We were considering setting up a large network server. After doing some number crunching, something became very very very clear: unless we were going to be moving to Gigabit Ethernet, 3 IDE disks in a RAID configuration were going to be more than sufficient to fill our 100MB LAN.
The point is, whether IDE will be "good enough" depends on what you're using it for. For a large fileserver, IDE RAID may well be good enough, depending on you local LAN. For video editting and other purposes where the data is used on the machine where the disks reside, SCSI's command queueing may be the better choice.
Unfortunately, that's not true.
Take a 7200 rpm SCSI drive. Take a 7200 rpm IDE drive. Rip off the electronics.
You now have two identical drives.
That's how it's been for most vendors for years now. SCSI does offer higher speeds (10 and 15k RPM), and the various other benefits spoken of, but reliability is not one of them. The electronics rarely fail on HD's. Instead it's a failure of a mechanical device (the motor, the heads, etc).
SCSI really doesn't serve much purpose on desktop machines anymore. Three times the cost for little or no performance gain. The days of IDE being vastly slower (even on the desktop) are gone, as are the days of IDE CD-R/RW's spitting out coasters if you as much as moved the mouse. There are a few people who will go out and buy the fastest SCSI drives out there, toss them in a RAID array, and then play games on it (no, I'm not kidding... a friend of mine did), but the cost-benefit there is so small as to be ludicrous.
...not to mention that I have yet to see a true IDE RAID controller. There's some nice RAID 0+1 controllers now that do OK, but RAID 5 still seems to be SCSI's domain.
Sure the question was is there any reason to use SCSI anymore in a desktop, for which my answer would be "there never was one for 99% of users" Of course gamer/benchmark freaks who need 200+ FPS (why?) will likely disagree with me.
This is simply not true for all SCSI busses. Each device will use the speed it is capable of. All devices are not forced to the slowest speed. It is true that slow devices may tie up the bus for longer periods than a time-sensitive device can tolerate, but then you shouldn't have placed the two onto the same bus anyway!
One thing that is true is that mixing single ended (SE) and low voltage differential (LVD) devices on a bus will cause all devices to behave as SE, with a possible lowering in the maximum speed possible for the LVD devices, but again, this does not necessarily mean they will all run at the same speed.
I have a dual P3-850 (was a P2-450). Under heavy CPU load it remains suprisingly responsive. However, if it's under heavy disk load, it crawls, even though Ultra-ATA isn't very heavy on CPU utilisation.
My previous machine was a single PPro-200 with SCSI disks. Under heavy CPU load, it crawled horribly. However, under heavy disk load, it remained much more responsive than my current system.
Therefore I conclude that SCSI really does perform better, even if the drives themselves are matched on throughput and access times. I think most benchmarks suffer a little from tunnel-vision and focus only on the raw disk performance without really taking into consideration what it all means in real world situations.
I put up with the worse overall performance of IDE because it's so much cheaper. Of course, I'm up to my limit (4 devices) and need a new controller if I want to add anymore. And, I have to remember to be careful about tying up the IDE bus attached to my CD-RW when I'm burning discs. I can't see the last point being a problem with SCSI.
I haven't seen any IDE controllers that sport a 64-bit/66 MHz PCI bus interface. SCSI already has PCI-X dual-channel U160/U320 controllers. Check out LSI Logic
IDE RAID is fine, it's cheap, but with newer IDE drives pushing 50 MB/sec (sustained) you could max out a standard PCI bus with three drives. Need more throughput? Then you're stuck waiting for PCI-X IDE RAID controllers, or at least 64-bit/66 MHz versions. And in the meantime, SCSI will just get faster.
Mainframe hard disks (historically SCSI) don't use remapped sectors. The drives are built to perfection. They are the top of the line. IDE drives are inferior, because the drives that have imperfections are sold off as IDE. This is because the bad sectors are remapped and hidden from the user.
Same old story people, OS/2 is the quality system but loses, Microsoft is the pile of junk yet sells to the masses. Likewise SCSI drives are the quality niche, IDE drives are the mass-marketed Microsoft-equivalent pile of junk with the bugs and flaws hidden.
While the manufacturing line at Seagate, IBM, Hitachi, Quantum, etc. take their SCSI drives *very* seriously, IDE is more like "yeah it's ok if we screw up a couple of sectors, couple of customers complain so what". The SCSI line failing is like Ford coming last at the Daytona, it's in a completely different league.
The reliability of SCSI drives outclasses that of IDE because the manufacturers discriminate during production (note IBM 120GXP fiasco did *not* affect its SCSI drives). If any cleanroom people can confirm my facts please reply.
A caveman dreams of being us, the incalculable power and riches. We dream of being Q, then what?
I've been a SCSI bigot since my Amiga days. Just 15 short years ago, all that was really available for consumer-level computers was SCSI, ESDI, and ST-506.
ST-506 was hardly an interface at all. You had to tell the BIOS the number of cylinders, heads, and sectors the drive had (sound familiar?), so that it could do the multiplication and convert logical block addresses into positioning information for the drive. You also had to enter the bad block list by hand, printed on a sticker affixed to the drive. An ST-506 interface was available for the Amiga-2000, and setting it up was predictably a bear.
SCSI saw its first consumer deployment on the Mac, and Amiga got it not too long after. No more CHS crap. No more typing in lists of bad blocks. All that intelligence was on the drive itself. Just plug the drive into the chain, tell the OS what SCSI address it had, and you were ready to start partitioning and using the drive.
So when it comes time for PCs to get intelligent drives, SCSI was the obvious choice. But no, they invent this new thing called IDE. What was different about it? As far as anyone could tell, the cable. You still had to feed CHS addresses at it; SCSI used LBA from the start. IDE drives from different manufacturers wouldn't work together; SCSI mandated interoperability. IDE now let you have two drives in your machine; SCSI already allowed up to seven.
IDE was touted as much cheaper, but it wasn't. SCSI and IDE drive prices were at near parity for years. Manufacturers were offering drives in both IDE and SCSI flavors (all other characteristics identical), with the SCSI flavor costing only ten dollars more (for a $600.00 drive, a typical price in those days, this was epsilon). It's only in the last few years or so that SCSI drive prices have skyrocketed for no readily discernable reason.
Add to that the fact that, even on a modern SCSI controller, all your old drives will still work. I have an old 600M 5-1/4-inch full-height Hewlett/Packard drive with a SCSI-I (asynchronous) interface. I plug it into the Adaptec AHA2940-U2W controller in my main rig, and Linux sees and mounts it just fine. Same with all my other old SCSI drives; I don't have to leave any of my data behind. It Just Works.
I also have an HP Omnibook 800CT laptop, which has SCSI built-in. All my drives work on that, too.
Apart from the artificially inflated costs, SCSI's only real headache is bus termination. But aside from that, the increased speed, flexibility, expandability, and reliability, for me, make SCSI an obvious choice.
Schwab
Editor, A1-AAA AmeriCaptions
I have some recent first hand experience with this one. I built a 980GB RAID out of eight 160GB ATA drives on a 3ware card, and I was all ready to ship it, and but I forgot that we needed to copy 500GB of our data onto it first. That was early yesterday. It is still copying files as I type this. It's pretty bad when it takes several days to copy files onto an array at full bore network speeds, just to fill it halfway up. Never before has network speed/storage space been at such a high ratio before.
I've had enough abrasive sigs. Kittens are cute and fuzzy.
The fastest IDE hard drives (WD's 7200 RPM 100 & 120 gig drives) aren't faster than high end SCSI drives but they're nipping at the heels of some very respectable SCSI drives. When you consider the price difference, it takes a very stubborn (or ignorant) mindset to blindly say "only use SCSI for servers".
There are some situations where good SCSI drives still have a distinct advantage. That's generally in the arena of database servers and systems with a high I/O load on the drives. But if you're not going to be building a server for that kind of environment, you're doing yourself (and/or your company) a disservice by not at least considering using IDE drives.
A 240 gig RAID with 6 channel IDE controller, hot-swap bays, and a spare drive can be had for under $1500 using the fastest IDE drives available. Capacity can be doubled for another $600. That's about half a terabyte of usable storage with the ability to suffer the loss of a drive and rebuild itself on the fly for a little over $2k. Try and build that for under $4000 with SCSI. Doubling the capacity of the SCSI system will cost $2800 (assuming you started off with 73 gig drives). More than quadruple the cost of expanding the IDE RAID. To use a PHB phrase, "what's the ROI on that SCSI system?"
Unfortunately, many consultants are unwilling to even consider using IDE. "Gee. I've never built a server with IDE drives before. Sounds like it could work but...SCSI IS FOR SERVERS! LONG LIVE SCSI!"
Look at the part numbers on the drives. They're the same.
Obviously this works better if you look at older drives, since there aren't many 7200 rpm SCSI drives manufactured still.
Sorry, but anyone thinking otherwise is trying to convince themselves that there's something magical about a physical transport medium that has the same performance requirements and characteristics.
They're also trying to convince themselves they're not being ripped off for buying SCSI.
I'm confused.. you say you have 2 drives, striped, and then talk about copying big files between them? IF they are striped they are one volume, and you can't copy things between them.
That was a mistype on my part, and what I meant to type was that while I have two 60GB drives off of a RAID controller, I haven't taken the plunge and striped them yet. As such they're both hanging as the single drive on their own bus, on two separate buses obviously.
I think the only reason IDE is more cpu intensive
It should be pointed out that while this is constantly restated, repetition doesn't count as evidence. It was ironic that just prior to seeing this debate, I saw this page which shows significantly higher CPU utilization for the two SCSI drives (mind you, they're extremely high performance drives, however they're not of a scale that would justify the difference between them and the IDEs). Each new time I replace my workstation I go through the whole IDE versus SCSI debate because I want to go with what's best (SCSI just has an air of superiority around it, much like Honda enthusiasts feel about their 115 lb-ft of torque VTEC engines : Enthusiasm, again, doesn't indicate that it's rational or based on any truths), but it seems that, firstly, it's extremely hard to find cold hard facts on the matter (i.e. basic metrics. Most of the evidence is anecdotal or based on uneven systems), but secondly that a lot of SCSI enthusiasts are very emotional about it. I have zero faith in anyone's personal opinion about the "feel" of one over the other: I remember back in the BBS days when a program made the rounds that promised to "convert your 386 to a 486!" and people would argue with me and ASSURE me that, yup, it made their system that much faster and smoother. A little persuasion and predisposition goes a long way when it comes to subjective measures, which is why I usually discount them.