· · Servers
· It’s days like these that make it fun working for Sun. The new server’s official name is the T5440; they call it a “mid-range” box, but to me it looks like a monster; count the numbers for cores, threads, RAM, and so on. It’s astounding what you can fit into a 4U box these days ... [2 comments]
· The Penguinistas like to brag about how GNU/Linux runs just fine on low-rent hardware, by contrast with competitors like Vista that need the latest gleaming iron to be useful. And they have a point; but only up to a point. I can testify from personal experience that an elderly 333-MHz Dell with a recent Debian totally sucks wind when you run WordPress. And the real point is, it ain’t operating systems that bog your computer down, it’s apps. [10 comments]
VirtualBox and Solaris
· Wow, we must have some pretty sharp M&A people. I’m no virtualization guru, but I thought that at least I knew about the serious players. But now we’re buying innotek, who make VirtualBox; I’d never heard of ’em but the reviews look good. I tried it out and learned a couple things ... [8 comments]
The T2 Servers
· These T5x20 servers we’re announcing today are a big deal. My bet is that they end up making Sun a lot of money; but on the way, they’re going to bring the whole server business (not just Sun’s piece of it) face to face with some real disruption ...
Testing the T5120
· This was going to be a Wide Finder Project progress report, but I ended up writing so much about the server that I’d better dedicate another fragment to the comparisons of all those implementations; especially since there are still lots more implementations to test. So this a hands-on report on a couple of more-or-less production T5120’s, the T2-based server that’s being announced today. Headlines: The chip is impressive but weird; astounding message-passing benchmark numbers; fighting the US DoD ... [1 comment]
· [Missed this one last week in Europe] We made a big splash earlier this summer at the High-Performance Computing show in Dresden; a concerted attempt to grab a bigger piece of the HPC market. I thought the coverage was pretty intelligent: the big deal about Constellation isn’t the blades or the disks, it’s the big switch. Anyone can stuff a bunch of racks with blades with hot chips on ’em; the hard part is getting them to work together, and whatever the approach, it’s easy to get bottlenecked on the messaging. The big new 3,456-way Infiniband switch is a brute-force assault on the problem, combining silicon, bus-ware, connector, and cabling wizardry: Josh Simons has close-ups and unveiling shots, but Jonathan has the best pictures. Why is this interesting? Josh has the numbers: 19% of the worldwide server market.
· We announced a bunch of blade stuff this morning and I honestly couldn’t think of anything useful to write about it, since it’s a decade or two since I was a sysadmin and discussion of I/O density and the like tends to go over my head. Apparently you can have Xeon and Opteron and SPARC and Windows and Solaris and Linux all in the same chassis, which sounds kind of terrifying to me. I did get to hear Andy Bechtolsheim give the spiel on the I/O design once, I understood maybe 20% but it sounded awesome. Anyhow, there is an interesting angle: it turns out we gave Scoble over at PodTech a pre-release exclusive for his video coverage. Which is not exactly high-gloss marketing, mostly just John Fowler in a monochrome meeting room with some journos, pulling dull silver blades and I/O modules out of dull silver chassis boxes and talking about the tech, taking a few questions. I think it’s immensely more effective than most product-launch bumph. As for the products, I don’t really have an informed opinion; but if you understand blades, you probably will if you watch the video.
Big, Big Iron
· Yesterday April 17th we did a great big honking announcement about servers, er make that an announcement about great big honking servers. Frankly, I know pretty well zilch about this class of machine; I can relate to Web pumps like the T1000/T2000, and to an x86 meat-grinder like the X4100 (which an irritated little bird working for one of our customers just told me that you can’t buy at the moment because they’re sold out); but these supertanker-scale things we cooked up with Fujitsu are from another planet. I went poking around blogs.sun.com looking for something hands-on. Bingo (but, er, anonymous bingo, which irritates me; I’ve expressed my opinion on this internally). First, “EXOTERIC” was down in the trenches with Fujitsu and the iron, see New Sun-Fujitsu SPARC Enterprise Systems. Second, “BM Seer” fires a volley of benchmarks: bang bang bang bang. I may not grok the finer points of mainframe technology but you can follow pointers from the announcement to the M9000 and look at the heavy-metal pictures. Ouf. [3 comments]
Rock Secret Contest
· Check out Jonathan’s piece today on the Rock. Sometimes he makes me nervous... but hey, let's have some fun with it. The first three people who can decipher his artful clues and guess (by commenting here) what the Big Cool Nifty Rock Secret is will win a round of blog applause from me and also a neat Sun sweatshirt or hoodie (my own, unopened conference loot sitting at the back of a closet, this is not a Sun contest etc etc) ... [12 comments]
· For some cheap chuckles, check the Web sites and consider the subtle differences between Intel’s spin and ours. I had no idea this was going to happen this week, but a few months ago I was in a room where Andy Bechtolsheim allowed that, yes, Intel was playing some damn decent catch-up, so the writing was pretty well on the wall. Andy’s explanation of the finer points of the trade-offs requires the best part part of two whiteboards to lay out and an alert well-caffeinated mind to follow, so I won’t even try. But... ain’t competition wonderful? For the foreseeable future, I bet almost all systems vendors will ship both Intel & AMD silicon. Hearty, honest congrats to Intel for getting back in the game. Both sides of the arrangement look to me like a win for both sides; as Otellini says, Solaris is a big deal in some markets that Intel wants more of, and if they say they can make it run better on the chips they build, I’d be inclined to believe them. Should be fun times in the server biz, hang on tight. [1 comment]
Thumper & Friends
· We announced a bunch of new boxes this morning (of, course, the damn Register has had the poop for weeks, I find our leakiness irritating). There’s a Real Big Opteron server (personally, I’m more of a scale-out than scale-up kinda guy, but big iron is a big part of our business). There’s a blade box. I know nothing about blades, never been near one. Then there’s the Thumper oops X4500, it’s interesting. I even have a grainy amateurish photo of the inside of a pre-production model ...
· This last week, I spent a bunch of time in a room with Andy Bechtolsheim. He brought along some of the new boxes he’s working on; some almost ready, some raw sheet-metal prototypes. I got some cool pix of a Thumper’s guts but they won’t let me run ’em. Anyhow, when you listen to Andy talk about these boxes, you realize that in their own way they’re beautiful. You have to watch it around him; if you ask a simple question like why one server’s I/O is surprisingly faster than another’s, you’re apt to get a 45-minute dissertation involving ten or fifteen pages’ worth of block diagrams (no, I’m not exaggerating). Anyhow, while they say the X64 boxes are our fastest-growing product line, I think that group’s feeling a little starved for love in among all the T2000 hoop-la. The marketing group sends me notes from time to time pointing to their news (for example a power calculator and a virtualization maniac), wondering if it’s bloggable. I keep telling them to grow their own bloggers, but they haven’t yet. Which is a pity, because what Andy’s doing makes a pretty compelling story.
· My thinking about computers changed this week. I’ve been watching the general trends in Web traffic and data sizes and so on, and been kind of worried that the servers aren’t going to be up to the loads that we’ll be throwing at them. Well... now I’m not worried about that any more. But we have some serious software pain coming at us.
Servers In the Right Places
· Earlier this month I lamented that we didn’t have much of a process for donating computers to projects that are doing good things. We seem to be making some progress on that, for example the T2000-tryout program seems to be running a lot smoother. But that’s not all; for example, an X2100 showed up Friday on the doorstep of Nexenta, as in GNU/Solaris. I think that this kind of thing is a complete no-brainer and hope that we manage to do more of it.
Oracle vs. Niagara
· Last week I was in Edmonton, and spent some time talking to the local Sun office and some customers. One of the things we talked about was our “throughput computing” product line, which is coming, uh I believe the party line is “late this year or early next year”. I like to talk about this stuff because in the Web-centric world where I live, a highly-parallel low-wattage machine hits a bunch of sweet spots at once. (Also, it presents interesting software problems.) Out there in the field, they seem to like it too; then on two separate occasions I heard “But we wouldn’t be able to use that.” I asked why, and they explained that Oracle’s idiotic per-core pricing formula would make it prohibitively expensive. Hey Oracle, Sun isn’t the only company that’s going to be shipping highly-parallel computers, and if there’s a technology out there that meets a lot of customer needs, and you’re standing in the way of them getting it, all you’re doing is moving the FYO point closer and closer.
By Tim Bray
I am an employee
of Amazon.com, but
the opinions expressed here
are my own, and no other party
necessarily agrees with them.
A full disclosure of my
professional interests is
on the author page.