On our new-ish Sun Fire T1000 server. 1 chip, 6 cores, 4 threads per core:

[mb]dlst1000v1 $ mpstat
CPU minf mjf xcal  intr ithr  csw icsw migr smtx  srw syscl  usr sys  wt idl
  0    0   0   22   214  110   10    0    0    0    0     7    0   0   0 100
  1    0   0    1     1    0    0    0    0    0    0     1    0   0   0 100
  2    0   0    1     1    0    1    0    0    0    0     1    0   0   0 100
  3    0   0    0     1    0    0    0    0    0    0     1    0   0   0 100
  4    0   0    1     6    2    2    0    0    0    0     2    0   0   0 100
  5    0   0    1     2    0    3    0    0    0    0     2    0   0   0 100
  6    0   0    1     2    0    2    0    0    0    0     1    0   0   0 100
  7    0   0    2     3    0    5    0    0    0    0     3    0   0   0 100
  8    0   0    2     3    0    5    0    0    0    0     2    0   0   0 100
  9    0   0    1     3    0    4    0    0    0    0     2    0   0   0 100
 10    0   0    3     7    5    2    0    0    2    0     3    0   0   0 100
 11    0   0    6     7    5    2    0    0    0    0     1    0   0   0 100
 12    0   0    1     2    0    3    0    0    0    0     2    0   0   0 100
 13    0   0    2     3    0    4    0    0    0    0     2    0   0   0 100
 14    0   0    1     2    0    1    0    0    0    0     2    0   0   0 100
 15    0   0    3     3    0    5    0    0    0    0     2    0   0   0 100
 16    0   0    1     2    0    3    0    0    0    0     3    0   0   0 100
 17    0   0    1     2    0    3    0    0    0    0     2    0   0   0 100
 18    0   0    1     2    0    2    0    0    0    0     2    0   0   0 100
 19    0   0    1     2    0    1    0    0    0    0     2    0   0   0 100
 20    0   0    1     8    4    2    0    0    0    0     2    0   0   0 100
 21    0   0    2     2    0    3    0    0    0    0     2    0   0   0 100
 22    0   0    2     6    2    4    0    0    0    0     2    0   0   0 100
 23    0   0    1     2    0    3    0    0    0    0     2    0   0   0 100

Lately I've been very interesting (some would say obsessed) with Solaris 10's Zones technology. It's a lightweight virtual machine technology, giving you some of the benefits of high-level VM servers like VMWare (isolation of services to specific operating system instances, consolidating apps to increase server utilization), with far less CPU and memory overhead. (Rather than running a bunch of separate operating systems w/ virtualized hardware, all of the Zones run off the same Solaris kernel.)

Michael Feldstein on the first round of court proceedings in the Blackboard vs. Desire2Learn patent suit: First 35 Claims of Blackboard’s Patent Ruled Invalid.

Finally today I started skimming through 37 Signals' book, Getting Real. An awesome, inspiring read for those stuck in the ditch on a project, over-planning and over-meeting.

you do so many things so stupidly. Your hardware is great! There's a lot of really awesome stuff about Solaris 10: Zones, ZFS, Dtrace, the whole deal. Even an acerbic BSD guy such as myself can appreciate it. But you have sooo much to learn.

For instance, take a look at this default port scan on a freshly-installed Sun running Solaris10. This machine was booted with the pre-installed OS -- I haven't touched a thing:

[mb]fcl46unt106519 $ nmap xxx

Starting Nmap 4.03 ( http://www.insecure.org/nmap/ ) at 2007-07-31 09:24 CDT
Host xxx.unt.edu (xxx.xxx.xxx.xxx) appears to be up ... good.
Interesting ports on xxx.unt.edu (xxx.xxx.xxx.xxx):
(The 1655 ports scanned but not shown below are in state: closed)
PORT      STATE SERVICE
21/tcp    open  ftp
22/tcp    open  ssh
23/tcp    open  telnet
25/tcp    open  smtp
79/tcp    open  finger
111/tcp   open  rpcbind
513/tcp   open  login
514/tcp   open  shell
587/tcp   open  submission
4045/tcp  open  lockd
6112/tcp  open  dtspc
7100/tcp  open  font-service
32771/tcp open  sometimes-rpc5
32772/tcp open  sometimes-rpc7
32773/tcp open  sometimes-rpc9
32774/tcp open  sometimes-rpc11
32777/tcp open  sometimes-rpc17
32778/tcp open  sometimes-rpc19
32779/tcp open  sometimes-rpc21

Nmap finished: 1 IP address (1 host up) scanned in 51.288 seconds

Sun, buddy: It's 2007. Do I really have to turn off telnetd, rlogin, et al, on a new machine? Why is this stuff even installed anymore by default?

The guy who pretty much changed my email life did a talk at Google couple of weeks ago, and now it's up. Behold:

This guy's blog keeps getting more and more interesting to me. Today he's taken on decision-making processes in IT. Quote:

I have often come into an organization where there is no clear understanding of their current business practices, policies, services, etc; that is, the business unit cannot “identify” their business processes. Many organizations actually choose to implement a technology with the hope that the new application/system will either provide uniformity in, or even create missing, business processes! This is completely backward!!!

Of the three platforms Blackboard supports for the Vista LMS (Windows/SQL Server on x86, Linux/Oracle on x86, Solaris/Oracle on SPARC), Solaris running on SPARC hardware has always been the preferable platform for us because 1) we already had an investment in SPARC hardware, 2) the higher costs (in licenses, administration overhead, training, migration) of Windows is a non-starter for us and 3) Java performance on Solaris smokes the other platforms.

The high (relative) cost of SPARC hardware is always the bear, and one of many reasons I'd been hoping that Vista would start supporting Solaris on x86 hardware (BEA does support WebLogic on Solaris x86). Sun's AMD and Operton-based servers are a class above other x86 platform offerings -- reading SmugMug's evaluation of Sun's AMD servers makes me kinda drool -- and we would, of course, love to spend less money.

Now, though, we have a couple of Sun's new-ish T1000 servers sitting in our machine room, just waiting to be installed. (These will be doing duty as Vista 4.2 application servers.)

The specs (hardware and financial) are impressive: a single CPU with up to 8 separate cores, and each core with 4 threads, for a base price of $4000. That's 32 simultaneous threads, in a single rackspace! Compared to even dual-core or quad-core UltraSPARC IV or AMD/Intel chips, the application density possibilities are staggering.

One way we measure price/performance is this: for Vista 3.0, the recommended spec is two CPUs and 4 GB RAM for each application node, where 1 node can support 250 simultaneous users. The T1000, having so much CPU and memory capactiy, can run multiple instances of Vista on a single physical box. Based on list prices (i.e. not what we actually pay), here's the breakdown comparing the Sun Fire V240 (our current mainstay) and the 6-core T1000 we've bought:

And frankly, that's on the conservative side -- my estimate is that we can get at least three nodes per 6-core T1000, bringing the price per 250 users below $2000. That beats the prices available on Linux and Windows on x86 hardware.

We also win by having fewer physical systems and operating systems to maintain. Putting more nodes on smaller boxes also means less server room real estate and less power consumption. All in all, it looks like a winning scenario.

Here's a really great piece by Michael Feldstein and Patrick Masson in eLearn Magazine, talking about the need for more flexibility in learning management systems. Quote:

Virtual classrooms should be more flexible than their physical counterparts rather than less so. Do you teach art history? Then you need an image annotation tool. But probably a different one than the image annotation tool needed to teach histology. Foreign language teachers may want voice discussion boards to check student accents. Writing teachers should have peer editing tools. History teachers should have interactive maps. And so on.

New York State University of Technology at Delhi CIO Patrick Masson discusses Wall Street Journal columnist Walter S. Mossberg's comments to more than 250 college presidents and other administrators attending the Chronicle Presidents Forum. Namely, that he said their departments are "the most regressive and poisonous force in technology today."