Links of the day

Okay, opensourcing now extends to processor technology in production: On OpenSparc.net the documentation of the UltraSPARC 2005 Architecture and the Hypervisor</a> were disclosed. Perhaps this leads to an Linux on UltraSPARC T1, but i have my doubts, that Linux will an equivalent level on perfomance on this processor when compared with Solaris.
CNN asks again Will Sun shine again?. The media is in desperate need of new puns regarding Sun. The sunrise/sunset-wordplays are overused.
Bigadmin has some very interesting articles on dtrace and grub. Worth a read. And for that matter there is an excellent article regarding <a href=”http://www.acmqueue.com/modules.php?name=Content&pa=showpage&pid=361>about obseravability in ACM Queue</a>. Mark Tremblay gave an interview to TechNewsWorld regarding future server processors:

Mostly by getting rid of power-hungry structures such as large associative memories, or content addressable memory (CAM) that are found in all out-or-order processors. Getting rid of any kind of speculation also means that the processor always does work that counts, as opposed to speculative work and/or predication.

Apropos future: Institutional investors seems to gain a little bit more conficence, that Sun Microsystems stocks will perform better in the future. TheStreet has an article about the potential in our stock.
Next big topic: Opensolaris gets XENified. Tim Marsland announces the availability of an Opensolaris capable to run on top of XEN. There are several posts on blogs.sun.com to celebrate this but this catched my eye: John Levon reports about the successful live migration of a solaris instance with XEN. One thing came into my mind at reading it: It would be really cool to have an cluster agent for Sun Cluster that migrates an paravirtualized solaris instance instead of an failover of the application by stop/start, at least for nonfatal errors and normal Clusterswitching for maintainance. Perhaps controlled by an master-instance that checks the health of the worker-instances and migrates them as nescessary …. extended with an steadily running replication of the system state to the standby system, this would be a hell of a cluster framework.
To close this entry: A news, that shocked me a little bit, as i worked for quite a time on an SGI Indy - SGI warns that bankruptcy might be year-end option. Okay, this happens when you believe analysts, that you should do commodity: Linux and Itanium. Nothing to differentiate, nothing to shine, nothing to make profit.