Slashdot Mirror

Most of my woes involved an onboard Adapter AIC-7880 (chip used on the popular AHA-2940UW) I had on an old motherboard. Similar problems happened with the AHA-2940UW PCI card version as well. The card treated the 68-pin Ultrawide SCSI and 50-pin SCSI as separate buses with sometimes confusing termination settings depending on what combination of connectors you used (internal 68-pin, internal 50-pin and external plug). There was also the issue of using 68-pin Ultrawide SCSI devices on "narrow" 50-pin buses. That setup required special "upper byte" terminators.

More on this mess here: http://www.scsita.org/aboutscsi/SCSI_Termination_Tutorial.html

SAS (along with SATA, USB 2.0, and Firewire in consumer level devices) has thankfully killed this mess off for the most part.

right now, i agree with you 100%, 10g is hardly needed anyhwere. but to address your orignal post, its that future thing that has me wondering. doubling up or tripling up on gigabit is currently standard practice for data intensive systems-- not just big iron, but rouge hackers cobbling together powerful clusters and grids for fun or profit. and it makes sense, its so damned cheap why not? besides, who can afford 10g? but with the throughput wars, how long can this really last? this is grassroots. this is mainstream, within the domain of computing hardware systems. for the actual computer industry, 1g is woefully insufficient.

grass-roots is people using drbd network replication with Xen to support live virtual host migration. if a filesystem fails, just migrate the hosts on that filesystem over to the host on the network backup and run them there. This sort of advanced system shuffling used to be the domain of blade systems and IBM, but now that bandwidth is becoming commodified and abundant we can start doing these things grassroots. Even now, some casual idiot can throw twenty one hard drives into a case for a couple terabytes of online storage. Soon with SAS (good overview), custom build storage will become only more of a reality.

Actually, SAS expanders use 4g infiniband interconnect. Maybe we just need cheaper infiniband. 4G is "nearly" enough.

So, currently storage solutions and blade systems are proprietary and expensive. With 10g and rapidly accelerating high availability and distributed systems, the linux kiddies are building it themselves.

this being said, i do wish to emphasize once more that I do really agree with you. there wasnt a single thing i didnt say yes yes and nod my head to in the parent.

scsi absolutely is not serial, duh

While he did screw up the second 'S' in SCSI, you cannnot seriously expect anyone who knows anything about the evolution of SCSI to take you seriously after you stated the above.

I will prove your statement false with a single counterexample: Serial Attached SCSI (PDF). Note the date of the document.

Remeber that with SCSI-3, the standard became more modularized in order to do things like separate the SCSI command set and the SCSI physical interface.

Here's the SAS FAQ from the SCSI trade association.

While this is a nice device, it's still bottlenecked by the PCI bus. Notice the speed on the specifications page - a true SATA drive can achieve this, and Ultra320 SCSI can shame it as long as you get a 64-bit PCI or PCI-X controller (like http://www1.shopping.com/xPF-IBM_ULTRA320_SCSI_CTR LR_2, although I suggest Adaptec instead of IBM). See http://www.scsita.org/aboutscsi/ultra320/faq.html for info on Ultra320. You get a lot more storage for the buck here. It's not solid-state, but a 73GB drive (15k RPM) -- see http://www.storagereview.com/articles/200209/20020 901ST373453LW_1.html -- can be had for under a grand.

Now... about large memory systems... what is the purpose of such a large device? Databases? Image manipulation? Analysis? If the PHB just told you to go out and buy something with more buzzwords, I'm sorry, but that's probably not going to solve whatever problem you have. What would the benefit be of having a single machine, as opposed to an expandable cluster of machines? Besides the cooling and electrical, anyway. If you have a legitimate need for such a large amount of memory in a SINGLE computer, you should definitely consider a Sun workstation or (more appropriately) a brand new Sun server.

On Serial Attached SCSI (SAS) systems. Each controller can address 16,256 devices. SAS is backwards compatible with SATA in that SATA drives can plug into SAS controllers.

There is actually a _great_ need to increase the communication speed between drives and controller.

For more information on SAS, see my Wikipedia article at:
http://en.wikipedia.org/wiki/Serial_Attached_SCSI

Since wikipedia is down or slow right now, here is the non-wiki version of the article and a link:
Serial Attached SCSI (also known as SAS) is a new generation of SCSI designed to allow for much higher speed data transfers. SAS does this by serial communication instead of the parallel method found in traditional SCSI devices.

SAS supports up to 16,256 addressable devices per port and point to point data transfer speeds up to 3Gbps, but is expected to reach 10Gbps by the year 2010. The SAS connector is much smaller than traditional parallel SCSI connectors allowing for small 2.5 inch drives.

The physical SAS connector is form factor compatible with SATA, allowing for much cheaper SATA drives to connect to a SAS backpane. SAS drives are not compatible on a SATA bus and have their physical connector keyed to prevent any plugging into a SATA backpane.

Serial Attached SCSI supports three transport protocols:

* Serial SCSI Protocol (SSP) - Supporting SAS disk drives
* Serial ATA Tunneling Protocol (STP) - Supporting SATA disks
* Serial Management Protocol (SMP) - Supporting SAS Expanders

SAS General Overview - http://www.scsita.org/aboutscsi/sas/tutorials/SAS_ General_overview_public.pdf

In this picture you can see that they are still using SCSI-style connectors. I thought the plan was to move to Serial-Attached-Scsi that used the SATA connector while still being SCSI. That way you could potential plug a SATA drive in the same slot as a SCSI was previous if your chipset supported it etc. It remember reading some Flash presentation on it. It was somewhere on the SCSI Trade Association website. They got a whole section on SAS on the site. So is it just theoretical? Anyone use it? Know anyone who does? Will it be out soon?

One drawback with SCSI is that you'll have to deal with a number of different cable/connector formats and terminators.

Dealing with terminators is easy: always install a separate active terminator block at the end of thew chain even if your device and/or cable claims to have one. Don't bother with passive ones.

As far as cable/connectors go, there are plenty of adaptors. Just learn the language.

I don't know much about SATA, so pardon the newbie question. But I've heard that SATA (Serial ATA) and/or SAS (Serial Attached SCSI) will make computers easier to upgrade, at comparable or lower cost, and is more scalable than the current (Parallel) ATA in use.

My question, then: in a SATA-based system, where legacy (P)ATA has been phased out entirely, would you be able to do other things AND burn a CD at the same time with little / zero impact on performance? Also, is SATA meant to replace AGP, too?

The artical leaves me asking why. This is the closest answer I got.

"IDC analyst Robert Grey said the ability to mix serial SCSI and ATA drives in servers and arrays has the potential to lower total costs of ownership for corporate users while also letting them customize storage setup to meet their needs."

OK, so SCSI costs more. But I see not a single technical reason for this other than economies of scale and the possible extra quality/longer warrenties that go into SCSI drives.

So why create a new SCSI? Didn't SATA take all the best things SCSI offered and added them to the ATA standard like queing? Is there any technical reason SATA can't add whatever SASCSI has? They added DMA ability to parallel ATA-33 which IMHO killed the biggest advantage SCSI had, I see no reason they can't come out with SATA 2004 and add whatever they needed for SASCSI instead of making a 2nd standard.

What does this mean? "Serial Attached SCSI complements Serial ATA by adding device addressing"

That's the only advantage SASCSI has over SATA I got when I read the FAQ.

The rest of the advantages seem to be "it lets you use SCSI drives, which everybody knows are more reliable and cost 10x more", but there is no reason for having SCSI drives. Just build better ATA drives!

320 megabytes is about 2.5 gigabits ... which is a lot closer to 3 gigabits than the erroneous 320 megabits figure.

No, the SCSI equivalent of SATA is called "SAS". Serial Attached SCSI.


http://www.scsita.org/sas/FAQ.html


What's really cool is that SAS and SATA share the same cabling and interface! SAS is a superset of SATA, that adds SCSI's features (multiple devices per port, and so on) to the basic one-device-per-port SATA design. The nice thing is that you can use cheap SATA drives on a SAS setup! This should be good for RAID. Think of SAS as "SATA Plus".

Here's a quote from the link above:

"Serial Attached SCSI complements Serial ATA by adding device addressing, and offers higher reliability and data availability services, along with logical SCSI compatibility. It will continue to enhance these metrics as the specification evolves, including increased device support and better cabling distances. Serial ATA is targeted at cost-sensitive non-mission-critical server and storage environments. Most importantly, these are complementary technologies based on a universal interconnect, where Serial Attached SCSI customers can choose to deploy cost-effective Serial ATA in a Serial Attached SCSI environment."

now, if you REALLY want ultra-fast disks in your desktop... firewire is FASTER than SCSI. up to 400 MB/s.

Firewire is 400Mbps, which is 50 MBps. That's faster than Ultra2 SCSI, but slower than Wide Ultra2, Ultra3 and Ultra160/320 SCSI. Check out this link for details. Firewire is still nice tech, and a fair bit smarter than USB2.0, but it's not the bandwidth king that SCSI is.

I am a systems engineer and work with million dollar machines.. AIX black raven... so. They use SSA drives

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

...if you're willing to live with the four-device limitation of Serial ATA...

Many of us simply can't live with the four-device limitation. SCSI and Fiber Channel shine when scalability is needed. There are many applications for multi-controller multi-drive RAID devices that ATA simply isn't cut out for. It is also nice to just be able to add another device when needed--SCSI is very convenient.

...eventually see the equivalent of ATA-300 and ATA-600 speeds, which far surpasses even the current Ultra-Wide SCSI 160

This is comparing something that doesn't exist to something that does. Also, Ultra320 SCSI is just around the corner.

Third, Serial ATA--unlike SCSI--doesn't require you to load device drivers out of the wazoo to support devices on the bus.

This is untrue. One SCSI driver allows me to connect any SCSI device: hard disks, ZIP drives, scanners, etc. The only additional drivers are those needed for non-SCSI devices, such as the parallel port or a modem.

...SCSI is still pretty expensive...

Not in the long-term. Good system administrators are more expensive than SCSI controllers, and the time and frustration saved more than pays for the SCSI controllers.

And in the home, SCSI really never had a foot-hold, so Serial ATA changes nothing.

In short, ATA never really competed with SCSI and never will. As long as ATA is crippled to be useful only in personal computers, it will never appear in big computers, multi-user computers, or high-performance workstations. These are not niche markets, either.

I just wanted to address two types of comments I've seen posted here:

* Encoding / decoding speeds are done at the speed of the medium. Encoding and decoding optical signals doesn't have any more overhead than PCI or IDE. The spec. writers and endec designers are well aware of these issues. That's why technologies like 10Gb Fibre Channel or Eithernet aren't ready yet -- not because we can't transmit at that speed, but that we can't build an entire NIC to sustain those speeds. (Give us some time: we'll be there soon enough.)

* Serial interfaces like Fibre Channel and Infiniband (and even Gigabit Eithernet) aren't replacing SCSI. They are replacing what you think of as SCSI: the 50 or 68-pin cable in your case. But SCSI is the protocol being used to talk to all those FC & Gig-E storage devices. SCSI over FC is called FCP (see T11's specs for more on FC). For Gig-E, most companies are looking into iSCSI, iFCP or FCIP (SCSI over IP or SCSI over FC over IP) for SAN-to-SAN communications. I forget the name of the spec for SCSI over Infiniband, but it pretty much rips it's ideas from the above specs. (sorry, no links for Gig-E and Infiniband at the moment: start at T10 or The SCSI Trade Association)

BTW, I refer to "serial interfaces" above instead of "optical interfaces" because a lot of this is actually copper. Most likely, Infiniband on the motherboard will be copper and off the motherboard it will be optical. Most of the Fibre Channel equipment I have isn't "fibre" but copper.