Slashdot Mirror
Is There a Place for a $500 Ethernet Card?
prostoalex writes "ComputerWorld magazine runs a story on Level 5 Networks, which emerged from the stealth startup status with its own brand of network cards and software called EtherFabric. The company claims they are reducing the load on the servers CPUs and improve the communications between the servers. And it's not vaporware: 'The EtherFabric software shipping Monday runs on the Linux kernel 2.4 and 2.6, with support for Windows and Unix coming in the first half of next year. High volume pricing is $295 for a two-port, 1GB-per-port EtherFabric network interface card and software, while low volume quantities start from $495.'"
Yes, there is a place for a $500 ethernet card, far, far away from this guy.
I wonder what has changed? I have never known the CPU to get dragged down by network traffic, but maybe in the network server markets it is different, However with the Ethernet chipsets being designed into the motherboard and integrated into the tight circle of RAM and CPU, it isn't clear there is a need for this.
How long before the network control is put into the CPU? It is going to be tough to beat that type of performance.
A most overlooked advantage to owning a computer is if they foul up there's no law against wacking them around a bit.
Is There a Place for a $500 Ethernet Card?
Of course there is, assuming the card performs as advertised. Sheer conjecture: the card likely has a lot of the smarts onboard. Maybe it has some of the TCP and IP stuff on board too (checksum, etc). Compare that to a crapbox $10.95 RealTek[a] card which generates interrupts like mad because it has no smarts and you'd probably be very suprised. (Think of comparing a decent hardware modem to a software based WinModem.)
[a] I had a sales-drone at Computer Boulevard here in Winnipeg just RAVE about RealTek cards. I said I really wanted 3 Intel or 3COM cards for a new work proxy server and he said 'Why? RealTeks are way cheaper and run at the same speed!' Retard.
Trolling is a art,
right inside my computer :)
No.
Move along please.
Short answer;
Where there are PHB's, there is overpriced hardware.
Million dollar PCs (sans gold-plating) and (quite seriously) mission-critical blade servers, customer ip routers, etc.... I have clients that pay upwards of $600 canadian now (though that's for quad cards with ample on-board processing to off-load from cpu horsepower).
if your internet connection is anything less than fiber, which is about 99.9% of all connections? Not to mention the fact that not many computers can actually handle that much data at once anyway.
They mention latency without saying what it is in the press release. I couldnt find it on the site. Maybe the tech docs have it. They compare it to Myrinet without saying what the latency is. It could be great. It could be crap.
They may even be integrated onto the motherboard and cost very little, relative to what we pay for standard onboard ethernet today. That is, if the technology is as good as advertised. If not, it will fade away from the memory of /.ers everwhere. The market will justify whether it is a quality product and worth the money. If it is, it will be mass-produced, competition will increase, and prices will fall!
This isn't exactly an entirely new concept. Intel have been selling their ethernet chips with built in SSL accelerators for quite some time, and the advantage of offloading duties from the software to the hardware (see Intel etherexpress vs RealTek style cards) is obvious. Whether these cards offload enough of the normal duties of a typical cluster node to be worthwhile should be interesting to see, as there are a wide variety of cluster load types and obviously these cards will have a niche to fit into alongside their competitors in the diverse set of demands around cluster networks. As for the price tag, I seem to remember gigabit cards being extremely expensive a few years back, and its probably pretty competitive with where they're aiming this product, alongside myrinet and infiniband.
Business Voyeur
I give Realtek 6 months tops to make thier own knock-off of the card for $24.95.
You say things that offend me and I can deal with it. Can you?
Frankly, now even moderately priced Ethernet card do at least some TCP Offload and buffering. I was surprised to see that the Marvell Yukon PCIe onboard chip on my new consumer motherboard did it.. Even on a seriously loaded server, I mean the kind of server that can output 1.8 Gbit/sec over a bonded link using samba, there isnt that much load on the CPU.
.. and latency mesured in microseconds, not milliseconds..)
This is the kind of product that is very cool, that was built by engineers without looking at the market needs. I would have no pity piting it against a good network card like an Intel, 3Com or Tigon3 and it would be VERY VERY hard to justify the price difference. At that kind of price, you can get myrinet which will get you a different kind of performance (2.4Gpbs per link.. up to 2 per card internally bonded
The Pentagon.
But not necessarily where the vendors think it is.
Back when I was working at a startup developing anti-DDoS technology, one of the biggest problems we were faced when implemented GigE, was the load on the PCI bus. (This was before we started using PCI-X).
It depends on exactly how customisable the network card software is, but if you could plonk a couple of those into whatever system you wanted - and if the cards themselves could do, say, signature detection of various flood types, or basic analysis of traffic trends then that is a very definite market.
I realise the core issue is not addressed (if your physical pipe is full, then you're fucked), but it takes the load of dropping the malicious packets off the host CPU so it can attempt to service whatever valid traffic actually gets through.
And then there is IP fragmentation. Bad fragments? Perhaps a dodgy fragmentation implementation in the stack? (you know which OS I mean) Lets just drop that before the host sees it and crashes.
I don't know, I can't find any real information describing what they do, but I can certainly see uses for this.
Wonder how if differs from Sun's Network Coprocessor, which used an onboard 16 MHz M68000 to offload TCP packet processing from the mighty 40 MHz SPARC processors in an SS690. Sounds like the Level 5 company's product (not Level 3, as the intro implies) also includes "improved" networking protocols that are supposed to be compatible.
Envy my 5 digit Slashdot User ID!
It's nice to see a piece of hardware that ships with linux drivers and promises Windows support later. So frequently applications and hardware are first supported under Windows and occasionally ported to other platforms.
The big place this will be beneficial is high performance computing solutions that currently use low latency application networks like Myrinet or Infiniband. By offloading the networking to the NIC, you free the CPU from having to handle the traffic and actually crunch numbers. Also, it allows you to use commodity gigabit switches rather than specialized, low volume Infiniband or Myrinet switches.
It would take a 4GHz processor to process the data involved in a 10Gb connection without this offload technology...THAT is why there is room for $500 network card.
Apple has fiber NICs as an option in the G5 Powermac, they are $500 and if none were noving, they likly would ditch them or lower the price...
So I wonder how this compares to this product by Bigfoot Networks. They won the moot corp contest, but it seems like someone beat them to it. They seem to want to make a network card like this, but for gamers. It seems like a similarly awful idea to me.
The name Level 5 refers to the network protocol stack where level 5 delivers data from the network to the application, according to Karr. The company isn't concerned about any potential confusion with Internet Protocol telecom Level 3 Communications Inc. On the contrary, he quipped, "It's working in our favor. People say, 'Yes, we've heard of you. You're a big company.'"
As lawyers at Level 3 begin salivating at thought of all of the potential lawsuits.
Back in the early-mid 80's (and probably even before then) IBM mainframes using SNA instead of TCPIP used special networking processors that handled all of that "networking stuff" so that the mainframe CPU (which really was a "unit" and not just a single chip) could just concentrate on running its jobs and not be interrupted by the communications end of things. Everything old is new again. Same situation, just smaller and faster (CPU and helper communications card take up 1U in a rack instead of 1 whole corner of the head end room).
We use Filers for storage at Gigabit speeds. Compared to our SAN/FC evironments, we see much higher CPU utilisation on our Sol 8 boxes, especially when attempting to get to Gigibit speeds.
Sorry, no. Intel and Broadcom have this sewn up. Why would we want to have to put ANOTHER network card driver through its paces?
InfiniBand is the answer for whatever problem this is attempting to solve.
$500 for a network card you have to have a good reason that you will need it. I am sure there are applications that will utilize it but for the price it may not be worth it. With sub $500 computers coming to age. It is probably cheaper just to split all your services onto smaller boxen and have a load balancing switch/router. Computers are cheap today $500 for a network card is steep and will only fill a niche market. Perhaps if the price was in the $50 range it would be more widely accepted. But with good enough systems at 1k and additional 500 could be used for a faster CPU other then a faster network CPU
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
How about they sell it for $100, when you buy the $600:year remote network admin package? Premium packages can include live security monitoring by companies like Counterpane. If you've got a need for network performance like that offered by these cards, you've got a need for sysadmin, including security. Or you've just got too much money.
--
make install -not war
The lower end of computers is going for ~300-400 dollars, from the latest Dull circular mailing I've recieved. How, preytell, is an ethernet card designed to "(reduce) the load on the servers CPUs and improve the communications between the servers" when you can just buy another server for the price of a few of these bad boys?
Windows has detected an undetectable error.
Level 5 Networks, which emerged from the stealth startup status with its own brand of network cards and software
I just saw a story on slashdot today that related to this.
This is all just part of a well-known cycle, with its own jargon file entry. In a few years they're going to be saying, "Hey, CPUs are so fast, it'd be cheaper to build a dumber network card and spend the money upgrading the CPU to compensate" and they're going to say that that's a new idea, too.
I highly doubt they're aiming these cards at the general public. The kind of folks who worry about this kind of performance aren't buying $500 computers, they're buying $5,000 + computers, and trying to tweak every ounce of performance out of them. I'm willing to bet my employer is going to look pretty seriously at these cards for some of our heavy-use systems.
Sometimes you can't "split all your services onto smaller boxen and have a load balancing switch/router". Not everything on the network is a web server.
Electricity is not free, esp in a datacenter. You should strike out for mid-grade boxen in most applications and be willing to pay a premium for anything that is going to be doing 24/7 traffic like a search engine. The idea of many smaller boxen being the best route is only true in university and research and developement type settings.
An Education is the Font of All Liberty
... and all your place are belong to us!!!
Etherfabric, aye? I wouldn't even wipe my butt with it.
Uh, huh. sure.
Your PC is not the target market. Clusters, large datacenters and applications that require communications as close to instantaneous as possible are. $500 is a drop in the bucket with the potential of a huge payback for those installations
Not everything is about Slashdotters home computers.
"I use a Mac because I'm just better than you are."
You're probably thinking of the i960-based cards, though Intel's PRO series adapters (not i960 based) do something similar (TCP checksumming is now builtin to the chipset and most OS drivers now know how to take advantage of that). That processor, and variants, were used in everything from network cards to RAID controllers.
They failed because the performance gain and CPU offload numbers were never enough to justify the price difference.
Ding ding ding. I forget who said it (maybe Alan Cox, but I'm REALLY not sure about that), but the opinion was along the lines that it would always be more benefitial to throw the money at a faster processor (or a second processor etc), because you'd get a performance boost everywhere. $300 buys quite a bit extra CPU horsepower these days, and there's no need for the hassles of custom drivers and such. Nowadays CPUs are just so damn fast, it's also not really necessary.
Please help metamoderate.
If this becomes popular, we're going to see exploits that pop the OS of the network card and get a bot or backdoor running on it.
Or better yet, tamper with the packets going to the system so they appear trusted.
I'm still waiting to see it happen to all the little bitty Cable/DSL routers.
When you can get eth0 to lie, it's all over.
My desktop p4 with a USRCourier gets lousy throughput compared to this pII softmodem laptop.
"A $500 LAN Card? Oh my God, Stevie, thats almost as much as my GeForce9900XTLSI+ cost!" Said the kid with the Lone Gunmen T-Shirt.*
"That's nothing, This 8-Track-ROM player off of ThinkGeekcost almost a cool grand" Stevie said, as the other nerds bowed around his glowing and chromed Frag Machine.
*Lone Gunmen T-Shirts coming soon. 8-Track-ROM's, too.
Looks good for your age..
Sure there is.
In a related story, the IRS has recently ruled that the cost of Windows upgrades can NOT be deducted as a gambling loss.
Take sun, some of their new server kit this year is going to ship 10Gbit/s ethernet on the board, which acording to their docs, is going to take 3 USIV procs to keep the bus saturated (6 cores). But when you are looking at 8 to 64 way server boxes, who cares about those 3 procs, especially when in 24-30 months it will take less than one proc to handle that load (Quad Cored + Moore's Law), and the eventually one thread will have the horsepower.
Surely those smart dudes at Via, AMD, Intel, Samsung, Nat Semi, and/or Motorolla aren't going to:
A) FUD this to death if it really works
B) File patent suit until doomsday to keep it locked up
C) Buy them out
D) Let them wither on the vine and then buy the IP.
09f911029d74e35bd84156c5635688c0
It's called a TOE card, and companies have been producing them commercially for a few years now.
I've had a hard time learning that there is a lot more to computers than a thousand dollar workstation or laptop.
There are a LOT of > USD 10K servers bought every year. If a USD 500 NIC can improve the total performance of such a server by 5%, then yeah it's worth it.
That being said, http://www.ammasso.com/ makes an Ethernet card (priced around $495, I believe) that utilizes both TCP offload and RDMA. The latency of the cards is around 10us. This is great for people needing a high-performance cluster, but can't afford the Infiniband interconnect.
That is, if the technology is as good as advertised. If not, it will fade away from the memory of /.ers everwhere.
/.ers that still go on about Blu-Ray discs.
Not necessarily true, there are
If I point out that you are incorrect, making me a foe does not make you any more correct.
If this card can do most of the work of IPSEC for me, it'd be a big win.
My main concern though is that with two ports, how can I be absolutely certain the packet has to go through my firewall rules before it can go anywhere?
Of course, the extra ports could be an advantage. If it could handle all the rules for you, then it might even be capable of functioning as a layer 4 switch and sending out a new IP packet before completely recieving said packet.
But, I'd want all the software on that card to be Open Source.
Need a Python, C++, Unix, Linux develop
Virtualization.
These are the kinds of NICs that would be put into a datacenter that is leaning heavily toward VMware GSX or ESX servers. Any bit of offload of the CPU in sharing the NICs is a good thing.
What?
What would be really cool is to have a programmable network card - so you could program it to recognise DDOS attacks, or signatures, or whatever. Kindof like taking your old Pentium 200Mhz that you run Linux on now as a router, and squeezing into a network-card size. I imagine something like that would have a huge market in the hacker/geek crowd, plus IDS companies would love it too (provided you can protect it from being re-progammed, maybe a jumper on it to turn off programmable mode or something).
Does anything like this currently exist?
Anyway, just my 2cents.
Sampizcat
Is it just me or does this news headline seem to be nothing more than a press release?
How about a $1300 rocket modem?
Unless you are google... they seem to be doing just fine with lots of smaller boxes running in clusters. Heck they don't even care about flakey hardware. Just because you can't see a use for something doesn't mean someone else can't.
Short answer no. Long answer some moron will buy it , but no one else.
I'll need one of those baby, ... oh my broadband upload only go up to 384Kbps? Forget it.
My CPU is still sleeping 99.9% of the time waiting for data over the internet
Hrm, you mean a router? ;)
- Adam L. Beberg - The Cosm Project - http://www.mithral.com/
Well, if it supports int13, you could use it to boot from iscsi. $500 network card compared to $1000+ fabric board that requires a seperate fabric/copper infrastructure... Don't know that it's a way I'd want to go with a server, but I can see some people thinking it's a Good Thing (tm)
I disagree with what you say, but I'll defend your right to say it to the death - Voltaire
I bet you can't get one for less than a grand. But if you buy one, it means you need it.
The programmable car would almost certainly cost more than a cheap upgrade from your 200 MHz pentium to a much faster MoBo/CPU.
I would wait on jumping to any type of conclusion and see what happens when EPSRC adopts this for ther 512 node cluster. If it really improves the performance by 10% it certainly should be something to look into. Now, is there is a market for it at end user's desktop? I don't think so. 9$ Realtek card would be just fine. :-)
"Smaller boxes" is relative. Google's cluster nodes are dual Xeons with terabyte+ HDs. For Google it is small, for anyone else, that is powerful computer you're going to be paying alot for. If you're buying one of those computers you're probably going to look at one of these cards, and that is exactly the market they're looking for.
I have a few servers right now that cost close to $30k each. (and this is at a 29 employee non-tech company).
$30k might seem like a lot to a Windows technician, but that is a cheap box for high-end Unix servers (won't even touch mainframe for that).
6 years ago, when I ran a network for a large company (in a very rural community, mind you) the most expensive server there cost over $320,000.
It was an AS/400 and it was/is a phenomenal piece of equipment.
Heck, it cost $900 just for 1 10/100 ethernet card back in 1999. (and PC nics were $15)
$30k is peanuts. Go checkout some 8-way IBM boxes. Or HP or high-end chompaq. And no, you do NOT reboot a $300k+ server during maintenance.
I looked at their benchmark web page http://www.level5networks.com/prod_etherfabricperf .htm where they claim that a typical PC with "conventional" ethernet burns 83.5% of CPU for communication overhead while only 16.5% remain to the application.
But they don't say which CPU was used - probably an 850 MHz Pentium III or something similar outdated.
Fact is, on a current 3.x GHz Pentium IV or an equivalent Athlon or Opteron the communication overhead is in one digit range, percentage-wise.
A famous computer science quote is:
"Lies, damned lies and benchmarks"
and another one is
"Don't trust any statistics that you haven't forged yourself."
Dedicated Linux servers (root access) $45 p.M.
Of course, in a Beowulf cluster.
HTH
threadeds blog
Thank you. I notice a tendency for people to immediately think about home/workstation class stuff when looking at new software or hardware. The problems are entirely different in the data-centre, which is where this product and TOE are aimed at.
Similar arguments can be had over the cost of storage. They say "Hey I can buy a 400GB hard-drive for $300 - why do you say it is going to cost you $500 per GB?"
"Well," I say, "I have to worry about MTBF, Seek times, IO channel bandwidth, backup, redundancy, vendor support etc. You dont."
Reminds me of the GS60's I worked on a few years ago. They were expensive enough that trading them in could buy housing for my whole extended family, and NICE housing at that.
And no, it wasn't a client with deep pockets and no common sense. Some software isn't architectured to work nicely on a distributed cluster, and people cannot wait for a solution to be developed. These GS60's paid for themselves in the first few months of deployment because they were used to reduce the over-generation margin of the power companies while still keep an acceptable safety margin for unexpected demand spikes.
Sure, another software architecture could be developed, but who's going to wait the two to three years for a prototype? And in some fields, they expect several years of proven technology before making the switch to "new" and less-tried solutions.
The bastards who rejected my idea for an Irwindale-based PS2/USB combo-card should be having second thoughts right about now.
You mean kind of like the Intel Network Processors?
m ily/
http://www.intel.com/design/network/products/npfa
If you have a machine (say on a machine running linux kernel 2.4.20-30.9smp) with a built in gig port (say with eth0 identified as eth0: Tigon3 [partno(BCM95704A6) rev 2003 PHY(5704)] (PCI:66MHz:64-bit) 10/100/1000BaseT) connected to a decent gigabit switch, and another machine (same card, same os)with a gigabit card, those two machines will achieve 940Mbps talking to each other (results via iperf, 0.0-10.0 sec 1.09 GBytes 940 Mbits/sec).
However, if you plug a windows box (2000 or xp, didn't have a 2003 handy) with either an add on card, OR built in gig (2000 vs xp) you get a rather less impressive figure of 550-630. Coincidentally, you'll get the same basic number if you run two instances of iperf on the same computer... This tells me the bottleneck isn't the PCI bus, it's the OS. If you can prove me wrong please do so...
Well, if it came to buying 4 of these network cards, or rewriting pieces of code to fit the new load balancing architecture, I'd buy the cards.
$2000 is nothing compared to extended development costs.
Not everything can be load balanced, and load balancing can incur more overhead if the protocol you are considering is a bad match for load balancing.
You can't load balance a database transaction, a backup file, or a streaming video feed. Don't assume that it's all HTTP / only HTTP.
All GBE cards are FC on the MAC layer. Get over it.
Here's where the problems come in:
1) buses suck. PCI-X is fast; a faster bus clock is better still
2) the problems for GBE NICs are, in no special order: dropping crap packets; cleaning up dirty cache (a huge problem for clusters, where this product is poised); session/protocol relationship managment; buffering up misrouted UDP; managing evil ports (setting them up and tearing them down); managing proxies and work arounds (a little SIP anyone? Burping up your IPSec???)
3) making the driver aware of what's going on so sessions don't vomit
4) not bothering the freaking CPU chipset every few milliseconds with useless crap
So, if they do any of these things, bless them and send me the bill. Because (save TOE cards), all of them hassle the drivers and chipsets to no end with stuff that could easily be offloaded. And to those that say, just put more cheapo load-balanced hardware on the job-- you're a chump and deserve to have stuff blown to bits when you mulitply failure points with junko doorstop hardware boxes with all of the brains of a goose.
---- Teach Peace. It's Cheaper Than War.
As CPUs get faster an interrupt costs you more in terms of lost CPU time. So, reducing the number of interrupts is more important now than ever before.
My 100 Mbs ethernet card generates about 5k interrupts / second when transferring data at about 30 Mbps. Gigabit cards are engineered to hold interrupts until a few packets of data come in so that a DMA can move a larger chunks of data. If this NIC reduces the use of interrupts even further (say by off boarding computation or even the entire TCP/IP stack and thus allows for even larger DMA transfers) the impact could be substantial.
Unfortunetly, my knowledge of computer innards stops here, so I can't calculate how much cpu time 5000 interrupts actually take or how the new PCI-Express bus changes interrupt processing or how much a benefit it would be to have say only 1000 interrupts / second instead of the 5000.
[a] I had a sales-drone at Computer Boulevard here in Winnipeg just RAVE about RealTek cards. I said I really wanted 3 Intel or 3COM cards for a new work proxy server and he said 'Why? RealTeks are way cheaper and run at the same speed!' Retard.
Computer Boulevard, as in what used to be Techtronics? I wouldn't take a single thing they say as fact, but that's just based on personal experience...
Interrupts are the one place where it's not remotely true. A faster processor will allow your system to handle significantly more interrupts. The whole interrupt model needs to be thrown out and replaced with something much better.
And while I'm at it, there are many cases where it's not true. Wherever you have a significant bottleneck, hardware acceleration helps tremendously. Tasks like encryption and (HighDef) video playback can max-out the highest-end systems available, while a $50 card can handle those tasks easily.
I don't think purpose-built hardware everywhere is the answer, but I do think having an FPIC/ASIC as a standard computer component could make for incredible speed improvements in most/all of the tasks that are hard for CPUs to perform.
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
A recent installation of our product had hardware specs something like two 32 processor IBM p690s, and a dozen smaller boxes (4 or 8 processors, I forget exactly).
Agreed, $30k is nothing in that market.
Yes. Blizzard needs these to run their World Of Warcraft servers.
http://www.myricom.com/ and http://www.myricom.com/news/050620a/
These babies used to cost $2500 just half a year ago. Now they are down to $800. Awesome.
Dedicated Linux servers (root access) $45 p.M.
Will it work with Windows 95 ? :)
and have a read of why the interrupt problem isn't a problem anymore, at least on Linux. Note the date too - October 2001.
LWN.net
NAPI has been implemented in the kernel.org kernels for a number of years now.
The Internet's nature is peer to peer - 20050301_cs_profs.pdf
I can't believe what I'm seeing here. The majority think this is a bad thing, it seems. I disagree.
I have a few Sun E450s in my shop and I am going to be moving them to gigabit ethernet soon. A gigabit ethernet card from Sun costs considerably more than this so it is an option as long as it will run on Sparc hardware with Solaris drivers. Sorry, but the Intel e1000 just isn't going to cut it here.
I'd like to see that article about 20 instructions (I assume these are ML instructions) handling an entire packet. This may be the case on CISC CPUs, but I just don't see it happening on RISC CPUs. I am not saying it's impossible, but I would definitely have to see that to believe it and I am genuinely interested in reading this article. Please let me know where I can find it.
I don't think some of you understand the difference between intelligent network chips and networkchips with a CPU core inside them. Take a look at Cisco's solutions. Their line is moving hastily towards ASICs. The idea here is that specialized hardware designed to perform a task will ALWAYS be faster at handling that task than a CPU running on the same clock. Cisco is proving this with Layer 3 switching vs routing. It's not clear what their solution is, but I'm willing to bet this NIC is an ASIC optimized for the purpose of handling TCP/IP traffic from an ethernet network. I have a hard time seeing how any CPU will be able to beat out an ASIC in this field today.
Also of note is memory and bus bandwidth. I have seen some comments about CPU usage and how it's negligible and what not. While I don't believe that either (I pay a lot for those cycles, I want to use as many for data processing as possible), I do believe that the CPU handling the TCP/IP stack takes a little more BUS bandwidth as well as memory bandwidth. If this is all handled on the card, both bandwidth usages will be reduced. Bus and memory bandwidth is already lagging way behind CPU speed as it is. It is my number one system performance limiter right now. The more I can eliminate it, the more productive I can be. Someone already mentioned large numbers of packets. This is a good argument as well. When dealing with large numbers of small packets, CPU usage on a CPU-based TCP/IP stack increases as opposed to a smaller number of larger packets. So in some cases, it depends on your network and it's configuration.
Also, consdier that maybe I do only get 10 more cycles per second from using this card. Is the card worth it? With CPU cycles at a premium and everyone here trying to purchase as many as possible and never a single idle CPU in any of my servers, I have to give a resounding YES! 10 cycles per second per CPU times the numebr of CPUs and seconds the NIC is in place is a LOT of cycles and most certainly worth $500 over the lifetime of a $30,000 machine. If they can prove it does what they claim on my hardware, count me in.
Is that reasoning applicable to RAID controllers too?
Congrations guys... you just admitted to causing actual trademark confusion... have fun in the courtroom.
they were used to reduce the over-generation margin of the power companies
By using up the excess power? :)
Move Sig. For great justice.
Damn straight there is a need for cards like this. Companies like Alacritech, Adaptec, Intel, and many others make very expensive offload cards for a reason. I realize that most of the readers here have got no experience outside of their mothers basement computer room but, if you don't know what you are talking about then please, stopping wasting bandwidth with your drivel.
Many high servers, high and low end benifit greatly from offload cards. These cards offload SSL, IPSec, iSCSI, and just the TCP stack and make high bandwidth servers fly. And yes, even a high end server by todays standard can have 30%-60% of a CPU chewed up by network traffic. iSCSI in particular will bury the fastest CPU if you have multiple gig interfaces running off a PCI-E bus and no offload engine in place.
It's not just TOE (tcp offload engines) that are coming along but full fledged iSCSI offload cards that offload the iSCSI stack as well as the tcp stack. Since the majority of hosts out there can't benefit from a $1000 2Gb fibre channel HBA, as they don't push that much data.
But an iSCSI connection might just be what the dr ordered. Plus have you priced a gigE port vs a FC port lately?
About 5 years ago I had to buy a new ethernet card for the AS/400 at work. It looks like a standard PCI ethernet card. The cost $AUS3000 ($US2300). So in comparison, this is cheap.
We also needed to get a new token ring card for an RS/6000. The IBM price was $1500 for that one, but someone tried taking their card out of their PC (an IBM card, still overpriced at $200) and it worked.
So yes, if you are dealing with the corporate world then there is a place for overpriced technology, regardless of whether there is any actual performance benefit.
>> Intel had the 586-driven smart-cards, and I believe 3Com had them as well.
>You're probably thinking of the i960-based cards
Actually Intel did have "586" named chip well before 486. Most people assume 586 means first Pentium processor, but they already used that number in late 80's. Never seen Intel ethernet card with one, but 3Com 3C505 aka Etherlink+ used it. 586 chip was ceramic version that got burning hot. 3C505 was supposed to be somekind of smart ethernet card and to be used with 3Com's own network OS. It never succeeded. There were drivers for Netware as well and that's what we used.
My condolences concerning your functional illiteracy.
It's been a long time.
Hate to say it but, imagine a beowolf cluster *with* these things
A blog about stuff.
Depends on what bus the built-in port is running off, normally they are attached to the same PCI bus as your expansion cards. If you are lucky they will be in a seperate PCI bus so at least you audio card want be fighting for bandwidth.
If you have an i875 chipset, the built-in NIC will be running over the CSA bus, which is a purpose built bus with more bandwidth.
Some server boards onboard chipsets use a PCI-X bas.
Hopefully this will all go away when PCI express takes over
Accelerating Ethernet in hardware, while remaining 100% compatible with the standard protocols on the wire, isn't all that new. Just over 2 years ago, I worked on a TOE (TCP offload engine) card at Adaptec.
l ymatrix.html?cat=%2fTechnology%2fNAC+Cards%2fNAC+C ards
http://www.adaptec.com/worldwide/product/prodfami
It was a complete TCP stack in hardware (with the exception of startup/teardown, which still was intentionally done in software, for purposes of security/accounting).
Once the TCP connection was established, the packets were completely handled in hardware, and the resulting TCP payload data was DMA'ed directly to the application's memory when a read request was made. Same thing in the other direction, for a write request. Very fast!
I'm not sure of the exact numbers but we reduced CPU utilization to around 10%-20% of what it was under a non-accelerated card, and were able to saturate the wire in both directions using only a 1.0Ghz CPU. This is something that was difficult to do, given the common rule of thumb that you need 1Mhz of CPU speed to handle every 1Mbit of data on the wire.
To make a long story short, it didn't sell, and I (among many others) was laid off.
The reason was mostly about price/performance: who would pay that much for just a gigabit ethernet card? The money that was spent on a TOE-accelerated network card would be better spent on a faster CPU in general, or a more specialized interconnect such as InfiniBand.
When 10Gb Ethernet becomes a reality, we will once again need TOE-accelerated network cards (since there are no 10GHz CPU's today, as we seem to have hit a wall at around 4Ghz). I'd keep my eye on Chelsio: of the Ethernet TOE vendors still standing, they seem to have a good product.
BTW, did you know that 10Gb Ethernet is basically "InfiniBand lite"? Take InfiniBand, drop the special upper-layer protocols so that it's just raw packets on the wire, treat that with the same semantics as Ethernet, and you have 10GbE. I can predict that Ethernet and InfiniBand will conceptually merge, sometime in the future. Maybe Ethernet will become a subset of InfiniBand, like SATA is a subset of SAS....
Dr. Demento On The 'Net!
It's simple really, if (s)he can get on with a little more hardware then that is surely so much cheaper and quicker than a redesign, build, test etc. etc.
threadeds blog
It depends on configuration. I agree with your's can't, but you can run those transactions, backups and feeds in parallel.
The web page for EtherFabric suggests that this approach passes a lot of or all the IP stack processing responsibility to the application.
Surely there are going to be a lot of security issues here? What will a malicious application be able to do with this level of access to the underlying services?
> The whole interrupt model needs to be thrown out and replaced with something much better. Any ideas? ;)
Me (Blog)
Way too expensive for a NIC.
Rather buy a no-name $10 NIC.
The best explanation of what they've done is a bit hidden away on their site:
http://www.level5networks.com/sol_approaches.htm
This has some block diagrams comparing conventional ethernet, other things like infiniband, and their technology.
The main reason why it's fast is that it reduces the number of context switches, since less work is done by the kernel, and reduces the number of times that the data is copied from one buffer to another.
These people have been working on this for some time and really know what they are doing. I'm suprised that it's as cheap as $500. If you need to accelerate a file server, web server, etc., then this is almost certainly a very cost-effective way of doing it.
And please be happy that they've released it for Linux *first*! This really shows how important Linux is to server applications.
There's a good case for making use of large ethernet packets, e.g. 9k instead of 1.5k. Huge numbers of small packets do have a significant effect on systems. I have benchmarks which show a doubling or tripling of throughput and significant reduction in CPU overhead on some cards by changing the packet size.
You do have to make sure all the machines on that particular LAN are using the same packet size and you're not routing the packets on to the general network, you'll see horrible performance problems if you do. So there's a case for a private backbone between high performance servers, used for inter server NFS, database queries, network backups etc.
Deleted
Could be the driver(s). Whether you consider that part of the OS or not is probably debateable (sp?).
Oracle, Sybase, Informix, MySQL, PostgreSQL etc.
The problem is that splitting an RDB out over 10 sub $500 machines *isn't* simple and isn't necessarily going to increase your performance, it is quite likely to reduce it.
It's a server network card and on a server, I/O is king, the CPU spends much of it's time waiting around for disk, then waiting around for the network. That isn't to say you can just grab all the CPU time to process ethernet packets, because you want the CPU available to run the application.
These types of cards are just as important today as they have ever been.
Deleted
cpu offloading makes sense if you have enough IO bus
capacity to drive many cards. With PCI, you can saturate one gige network, and half of another. In a supercomputer from a few years ago, they used to use two slot PCI buses (had dozens of buses.) That would be a great place of TCP offloading, but most people arent going to notice.
You need for the IO-bus to be vastly (say 10x or 100x) faster than the networking bus for it to be worthwhile, say to use SCSI-over-IP and get several Gbytes/second in real throughput over multiple channels. Commodity servers are stuck with a couple of hundred megabytes per bus, which cpus can handle. There isnt any point to scale to.
Worse, Infiniband is going to make inside and outside buses, networking and storage technologies, go pretty much at the same speed.
Just find a way to promote it among the audiophiles.
Lisp is the Tengwar of programming languages.
Alacritec is/was making such accelerated adapters so the answer to this is yes, there is a market for them since there already is a competitor.
However their website seems to have gone dead so maybe the ultimate answer is: No there is no place in the IT market for accelerated network cards.
- Erecting yet another edifice brings on the huge and unavoidable overheads of yet another different CPU instruction set, yet another real-time scheduler, another code base, another set of performance and timing bottlenecks. Another group of programmers. Another set of in-circuit emulators, debugging tools, and system kernel. Another cycle of testing, bug fixes, updates.
- It sets up a split in the programming team-- there's now much more reason for finger-pointing and argument and mistrust.
- The extra money would usually buy you another CPU and lots of RAM, resources that would benefit every part of the system, not just the network I/O.
- The separate I/O processor usually requires the geekiest and least communicative of the programmers-- not a good thing. The manuals for the I/O card are likely to be very brief and sketchy, and rarely up to date.
- The I/O processor is almost always at least one generation of silicon technology older than the CPU, so even though the glossy brochures just drip with Speeeed! and Vrooom!-y adjectives, it's not that speedy in comparison to the CPU.
For examples, see the $4000 graphics co-processor that IBM tried to sell for the PC (IIRC the CPU could outdo-it). The various disk-compression cards for the PC. Also see the serial ports on the Mac IIvx (very expensive and not noticeably better). Don't forget the P-code chip for the PDP-11/03. All very expensive and blase' performance/$.These would be a good place for these, either with embedded slower x86 or non-x86 systems, Via C-3 processors, etc. It could provide an excellent way to reduce network bottlenecks and retain low environmental costs. Those are things that throwing money at faster CPUs can't touch.
If I'm going to spend $38,000 on a Sun V40Z, soon to have 4 dual-core Opterons, 16 GB RAM, dual fast SCSI disks, wouldn't I spend another $500 to keep it busy? If the product really works, it would be cost effective for squeezing a dollop of additional performance out of mid-to-high end "commodity" servers ($5-50K). Not everything is a $2k white box server or desktop. There are a lot of applications (Oracle, Apache, third party app servers) where it is already cost effective to buy the fastest processors and lots of memory. Paying a 10% premium (or less) for better network performance is easy to justify.
Envy my 5 digit Slashdot User ID!
See subject... -G
I've been buying gigabit ethernet cards for $25 each. They're not as bad as the first generation. Beyond $100 or so you might be wasting your money.
I just think there could be some truth with this smart cards.
It may be similiar to SCSI against ATA/IDE.
I have some older workstation with an 3Com 3cSRV97 Thyphoon smart NIC and it is quite fast despite its age.
Before some days I did a dd over netcat and the load was 2.65 IDE/Realtek Crap against 0.25 15K SCSI/3c97SRV NIC.
Maybe there ist not all about CPU speed here, despite most people are only looking for this.
Well, that depends. Does it support the following?
- 256MB DDR2 memory
- 16 Unified Pixel and Vertex Shaders
- OpenGL 2 and DirectX 9
- SLI mode for extreme performance
If not...then nope, I will stick with something cheaper!Unstable Apps: Our Android Apps Don't Suck
It surprises me how many people on slashdot either cannot or will not read. From the official website:m
f .htm ] seem to indicate just this.
http://www.level5networks.com/prod_etherfabric.ht
"...
In the conventional network I/O architecture, applications make Sockets or MPI function calls to interact with the network, and the Sockets library software converts these to system calls into the operating system kernel TCP/IP stack. With the EtherFabric architecture, Sockets and MPI function calls from applications invoke EtherFabric library routines that perform the necessary TCP/IP protocol processing in the user, or application, space. They move data directly to or from the virtual hardware interface that has been assigned to the specific user context. This approach eliminates context switching and greatly improves cache locality for both instructions and data, producing substantial reductions in latency and improvements in CPU efficiency.
Furthermore, EtherFabric hardware notifies each user process about incoming packets by placing events on the event queue for the relevant virtual interface. EtherFabric software de-queues these events and performs protocol processing on incoming packets without the need for interrupts. Substantial further savings in CPU cycles are made by avoiding the inefficiencies of interrupt processing"
What they are saying here is that ALL TCP/IP Processing is done in USER-SPACE. From what the blurb says I would assume the actual stack is in user-space and that only interaction with the kernel would be on connection setup/teardown. They don't say how they actually move data in and out of the application but I would bet my bottom dollar they probably do memory mapping or some other trick to allow data movement in and out of the application without corssing the kernel/user application boundary.
The actual design seems like a combination of a very smart ethernet NIC and a kick-ass user-space application space IP stack. I can see a lot of advantages with this approach
1. User-space network processing means more efficient resource usage and accounting
2. User-space network processing means more efficient data transfer between application and stack.
3. By doing the network processing in user-space on the host as opposed to a typical TOE, you are not constrained by the NIC. A faster CPU means you benefit from faster TCP/IP processing.
4. This also makes upgrades and software bugfixes much easier and more maintainable -- replace a library as opposed to having to replace the card or firmware of the card.
5. By using an efficient NIC with hardware support for virtual channels you have very efficient demultiplexing, you dont waste cycles in software oding this.
Overall, it seems to have a lot of benefits --
a)Because it's Ethernet based no need to rip out existing infrastructure
b)Because it's normal tcp/ip as opposed to some weirdo network protocol it's well understood and compatible with existing technology
c)because it's a drop in solution no need for application or kernel redesign and re-engineering.
IMHO, if this does what it says it does, it will revolutionise traditonal ethernet cluster/data centre processing. It's analagous to a turbo-booster on a car, just switch it on and hold on to your hat. Their benchmarks [ http://www.level5networks.com/prod_etherfabricper
As faster networking technologies become more prevalent (eg 10 Gig Ethernet) the problem of network cpu consumption will become a bigger and bigger issue for server type applications. The toss up will be between this approach and the TOE approach -- coupled with the advent of multi-core CPU's I can see this being a much more flexbile solution than TOE's. However, time will tell.
It always seems to me that so many people who comment on Slashdot have an incredibly narrow view of the IT industry. A product like this is not intended for the home user. This is an enterprise product, designed for use in compute environments where you are shipping large volumes of data amongst computers, where removing the CPU load from the system free's up cycles for the system to do the real work it is intended for, and keep those CPUs supplied with enough data, so you're not waiting on IO. Enterprise adapters demand a significantly higher price than the $25 card you're putting into your uber gaming machine. When you're talking about systems with 20+ CPU's and 40+ gigs of RAM, you're in a different league altogether.
It surprises me how many people on slashdot either cannot or will not read. From the official website:m [level5networks.com]
f .htm [level5networks.com] ] seem to indicate just this.
http://www.level5networks.com/prod_etherfabric.ht
"...
In the conventional network I/O architecture, applications make Sockets or MPI function calls to interact with the network, and the Sockets library software converts these to system calls into the operating system kernel TCP/IP stack. With the EtherFabric architecture, Sockets and MPI function calls from applications invoke EtherFabric library routines that perform the necessary TCP/IP protocol processing in the user, or application, space. They move data directly to or from the virtual hardware interface that has been assigned to the specific user context. This approach eliminates context switching and greatly improves cache locality for both instructions and data, producing substantial reductions in latency and improvements in CPU efficiency.
Furthermore, EtherFabric hardware notifies each user process about incoming packets by placing events on the event queue for the relevant virtual interface. EtherFabric software de-queues these events and performs protocol processing on incoming packets without the need for interrupts. Substantial further savings in CPU cycles are made by avoiding the inefficiencies of interrupt processing"
What they are saying here is that ALL TCP/IP Processing is done in USER-SPACE. From what the blurb says I would assume the actual stack is in user-space and that only interaction with the kernel would be on connection setup/teardown. They don't say how they actually move data in and out of the application but I would bet my bottom dollar they probably do memory mapping or some other trick to allow data movement in and out of the application without corssing the kernel/user application boundary.
The actual design seems like a combination of a very smart ethernet NIC and a kick-ass user-space application space IP stack. I can see a lot of advantages with this approach
1. User-space network processing means more efficient resource usage and accounting
2. User-space network processing means more efficient data transfer between application and stack.
3. By doing the network processing in user-space on the host as opposed to a typical TOE, you are not constrained by the NIC. A faster CPU means you benefit from faster TCP/IP processing.
4. This also makes upgrades and software bugfixes much easier and more maintainable -- replace a library as opposed to having to replace the card or firmware of the card.
5. By using an efficient NIC with hardware support for virtual channels you have very efficient demultiplexing, you dont waste cycles in software oding this.
Overall, it seems to have a lot of benefits --
a)Because it's Ethernet based no need to rip out existing infrastructure
b)Because it's normal tcp/ip as opposed to some weirdo network protocol it's well understood and compatible with existing technology
c)because it's a drop in solution no need for application or kernel redesign and re-engineering.
IMHO, if this does what it says it does, it will revolutionise traditonal ethernet cluster/data centre processing. It's analagous to a turbo-booster on a car, just switch it on and hold on to your hat. Their benchmarks [ http://www.level5networks.com/prod_etherfabricper
As faster networking technologies become more prevalent (eg 10 Gig Ethernet) the problem of network cpu consumption will become a bigger and bigger issue for server type applications. The toss up will be between this approach and the TOE approach -- coupled with the advent of multi-core CPU's I can see this being a much more flexbile solution than TOE's. However, time will tell.
So, you've developed a fine product, produced a few of them, and now you want to find "a place for it". This is called "marketing". My name is Mike, and I'll be your "marketing consultant" today.
No,I don't actually produce anything, and all of your techno-babble is hurting my ears. Give me a one-sentence summary of your product, and sign this simple contract granting me 60% of the gross proceeds. Then I'll wave my magic marketing wand and generate an absurd list of overinflated or technically-impossible claims vaguely based on your one-sentence summary. Then I'll whip off some glossy copy (I'll need some up-front budget for that), and hang out with the lovely models we'll hire to promote your product. . . .
Yes sir, that's how it's played. I'm Mike, are you ready to find your market?
You meant the serial ports on the IIfx, not IIvx. They were 6502s that babysat the ports. Yeah, they pretty much sucked. They were worse at most things. They were actually better at receiving LocalTalk packets with virtual memory turned on than the standard ports though. But honestly, if you paid that much for your computer, you probably had an ethernet card.
My understanding was the PGA actually accelerated autoCAD, which was enough for some people. Overall, it was a huge flop.
Apple also had some video accelerators. The 8*24GC was their first one, it was rarely faster than the regular video card* and vastly less compatible.
*The 8*24GC was faster for line-drawing operations, and also somewhat helped overcome the slow NuBus in the IIfx. But any decent onboard video (like the IIci and later) was much faster at all standard cases.
On the topic, Apple also had a smart ethernet card (it used A/ROSE). It was also slower than a non-smart card. Apple released a "dumb" driver for it at some point that bypassed the coprocessor and was faster than the "smart" driver. They then released a card with all the smarts stripped off. It was much cheaper.
I while I'm posting, I'm gonna say that I think that there is a market for smart NICs, for certain operations. I don't need one, but packet filters/forwarders definitely find them useful.
http://lkml.org/lkml/2005/8/20/95
The white-paper and the web pages on the company site which describe the implementation talk about how this is done.
This card, and the software which drives it, differ from traditional ethernet accellerator cards and from alternative network protocols (like myrinet and iWarp) in several ways.
Alternative protocols not only require using a different software API but also require custom hardware at both communication endpoints.
Traditional hardware TCP/IP accelerators run the bottom half of the stack in custom silicon. This does tend to help reduce host CPU load but suffer from a number of problems. Since host CPU speeds have tended to increase regularly, they often helped for only a brief period of time. They also tended to help most for large packets but helped little or not at all for small packets.
This technology claims to help large and small packets equally well, and also claims to reduce packet latency across the board. It does so by running the bulk of the TCP/IP stack in user space rather than via system calls. The hardware runs ethernet Rx and Tx processing but does not implement the higher level IP protocol processing. Instead, once connections are established, the ethernet frames coming from the hardware, are fed via a system call interface to the application process to which they belong. Then, no further context switching between kernel and the process are required. The top end of the hardware driver and all of the subsequent IP layers, are executed in the context of the user space process. They are linked to the app via shared libraries.
Basically, instead of the linking the IP calls against code which requires frequent switching between user and kernel space, the entire upper half of the stack is run by the application sending and receiving the packets. This offers uniform benefits in packet latency across all packet sizes, and offers improvement in throughput as well.
I assume that all that is required is to link against a different set of shared libraries to gain these benefits (and of course to have the custom hardware on at least one side of the comm. link). This looks very good in principle.
The following page provides an overview of the technology and compares it to each of the competing mechanisms.
http://www.level5networks.com/sol_approaches.htm
Sheesh... prices are just getting out of control.
Now where did I put those 30pins SIMMs... time for an upgrade to 20M!!! Whoo hoo... I know sounds extravagent for somebody who is against $500 ethernet cards. So maybe I'm just being hypocritical.
Where is your middle tier? Where is your redundancy? And you gave direct db access to web servers that on are on the public internet? So when your web server gets owned, the hacker has full access to the DB?
Every bit of the handshake, ... is going to wrap up that much faster if things are faster
Look up connection pooling.
where every page renders around a hundred or more queries
Wow. I'm lost for words. Seriously though, how do you have a job? This has to be the saddest architecture i've ever seen.
Have you ever been to a turkish prison?
... named "Level 7"? The abbreviation would be L7. I guess that would just be too square, eh?
Of course, if everything used MSI or other "smarter" interrupts, things would be a little better, but the whole context switch problem is still costly regardless.
"It's tough to be bilingual when you get hit in the head."
How can Timothy and the poster possibly feel entitled to comment on networking news when they cannot even distinguish between 1Gb ( GigaBIT, what this card is ) and 1GB ( GigaByte, what they wrote ).
Stick to posting Apple Ads guys.
Well, if Intel was to grant me a few million dollars for research, I'm sure I could come up with some ideal system within 12 months...
Off the top of my head, I'd say some type of hybrid approach, using both polling and interrupts would reduce the problem immensely.
Alternatively, you could stick to the old tried and true method of having a seperate chip in the system that does nothing else but service interrupts.
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
From a market perspective, there's often a period when it's more efficient to build new capabilities into a fancy expensive coprocessor than to fix the underlying bottlenecks. Sometimes you can start shipping boards before the problem goes away, sometimes you even sell enough to make tons of money, but sometimes only the top 1% of computers in the market are really affected so even if you sell it to everybody you don't make back your development costs. And sometimes people would buy the fancy cards for a while even if there were better approaches - back when Van Jacobsen was getting a 3 MIPS Sun 3/60's Ethernet to push 8 Mbps of FTP traffic by fixing the TCP and IP stacks, faster Intel 80386 and 80486 machines generally needed accelerator cards to get 4-5 Mbps instead of 1-2 Mbps, much of which seemed to be related to intricacies of DMA and interrupt handling on the PC as well as on slow bus speeds. Of course, the prices have changed by a few orders of magnitude - it was hard to make money selling $1000 cards when the bottom end of the market was $500, but it's a lot harder when the bottom end of the market is $19 (for gigE, or $3 for 10/100) and the moderately-expensive cards cost $79.
The grandparent article was talking about ATM processor cards, so I'd guess the timeframe is about 5-7 years ago when some people still believed in ATM to the desktop. It looks like my online copies of Fore manuals are from 1999, and when we bought the switch in my lab, the vendor threw in a couple of PCI-based ATM cards which could ostensibly work in both a PC and our blazingly fast Sparcstation 5, though we never did try it in the Sparcstation. The card had some critical ATM features implemented in hardware - segmentation/reassembly between AAL-5 data frames and the layer 2 53-byte cells, priority queuing for different types of ATM PVCs, and ABR flow control. It probably also did AAL-5 frame headers and checksums, and it wouldn't surprise me if it also did UDP checksums just because it had hardware around. Ethernet cards in those days had reputations ranging from deadly stupid to adequately smart to way too clever to be actually useful, pretty much as they do today.
These tradeoffs happen in lots of other parts of the architecture - people like Henry Spencer used to argue in favor of dumb-framebuffer graphics systems with CPUs doing the heavy lifting, but most of the higher-end graphics cards today use the card for SIMD crunching on features like shading and textures, with an AGP*N channel into the CPU because PCI is too slow, and the CPU-vs-card arguments are mainly for geometry calculations (I think CPU is mostly winning?) Serial I/O, to the extent anybody still uses it, seems to have mostly moved into the Southbridge chips, but before that you could get a major performance win by using a 16550A UART instead of the dumber UARTs, and of course I remember the Sys3 / Sys5 folks using Unibus KMC11s to service the DZ11 I/O processors for functions like cooked-mode I/O rather to avoid bothering the CPU (I don't remember the BSD folks ever adopting that, but I don't remember why at this point; perhaps emacs made cooked-mode uncommon enough that it wasn't worth the bother, or maybe the new tty driver could get more work out of a dz11.)
One other place where there's an ongoing argument between using the CPU vs. an I/O board is modems - the Windows side thinks it makes much more sense to do most of the work in the CPU, and use a dumb modem with a fast PCI interface, while the Linux fo
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
It's certainly not my best post. But I don't think it's quite that bad.
Need a Python, C++, Unix, Linux develop