By User12608550-Oracle on Jul 04, 2016
If you're there, visit the Oracle booth, and attend our presentation on M7 performance and Graph Analysis by me, Marty Itzkowitz, and Ruud van der Pas, Courtyard Marriott, Rio Grand A Conference Room, Tuesday 10:30-12:30.
Special thanks to colleagues/coauthors Glynn Foster and Joerg Moellenkamp
and tech reviewer Tom Plunkett for this effort!
The earlier Solaris 11 book, Solaris 11 System Administration the Complete Reference, still relevant and available
at the San Jose Cloud Day, a one-day public cloud forum, you will have an opportunity to:
And for some truly amazing M7 SPARC processor performance and security features, check out our Software in Silicon Cloud.
Well, here we have an article on containers, Are Containers the Beginning of the End of Virtual Machines?, and NOT ONE MENTION of Solaris containers (or "zones", if you prefer), that have been in production for nearly 10 years. I can only assume that the author had to TRY VERY HARD to avoid any mention of this commonly used virtualization technology!
I would like to focus on one small but important new feature that allows you greater choice in Solaris OS virtualization: Kernel Zones.
You create and manage Kernel Zones much like other zones the way you did it on Solaris 10 or earlier Solaris 11 releases. But now on 11.2 you can run different Solaris 11 kernel versions within non-global zones, similar to the earlier branded zones idea for Solaris 8 and 9. Solaris 11.2 now provides a platform for Type II hypervisor like support for different Solaris kernels.
This means that you now have two choices for virtualizing multiple Solaris kernels on the same physical server. Use Logical Domains (LDoms) on the T- and M- servers to provide fully separated Solaris kernels running directly on allocated SPARC cores, or use lighter weight kernel zones hosted on a single Solaris 11.2 global zone. And this provides intriguing variations and combinations of virtualization-within-virtualization: zones within LDoms, for maximum flexibility of application containment and deployment.
Check out Marcus Flierl's video on Solaris 11.2 virtualization here, and start thinking about all the new ways to deploy test, dev, and production applications in virtualized Solaris environments.
Starting long ago with von Neumann's "The Computer and the Brain" , computing analogies have both illuminated and confused many important IT concepts. The use of the term "memory" for persistent data storage, for example has a fairly clear and beneficial connection to human recall concepts. Less clearly helpful are analogies to thinking and consciousness common in AI research, as well as the current discussions around "cloud" computing. The corresponding characteristics of meteorological clouds and modern distributed computing are imprecise and misleading. It’s not clear how that analogy helps understand the critical technical concepts.
Now we have yet another unhelpful analogy, “Fog Computing” . Attempting to characterize the “Internet of Things”  as an all-pervasive, obscuring “mist” explains nothing about the nature of ubiquitous, embedded computing services. And of course, computing vendors will jump at the chance to exploit the latest analogical buzzword to promote their products .
It’s difficult to communicate anything without using analogies,
since that’s how the human brain works — we think using analogies
. But we should be careful in selecting the source analogs when
trying to explain complex concepts. Poorly chosen sources can
confuse and limit thinking and can hinder solution development.
Surely there are better source analogs than clouds and fog. The
“web” is clearly better than the “cloud” in conveying the idea of
connectivity. But what is the ideal analogy for computing services
that will eventually fill every corner of our daily lives, using
our always-connected devices like smartphones and tablets, and the
embedded services in our homes, cars, businesses, and social media?
I don’t think it’s atmospheric phenomena. And I’m not sure it’s the
“invisible computer” analogy either . It’s probably more like
oxygen , although that implies that we can’t live without it.
Hmmm…maybe we can’t.
If you want to play around with your own Hadoop-based cloud, be sure to check out these How-To articles on the Oracle Technology Network:
Hey! You! Get onto my cloud!
With an amazing 1,536 threads in an Oracle M5-32 system, the number of threads in a single system has never been so high. This offers a tremendous processing capacity, but one may wonder how to make optimal use of all these resources.
In this technical white paper, we explain how the heavily threaded Oracle T5 and M5 servers can be deployed to efficiently consolidate and manage workloads using virtualization through Oracle Solaris Zones, Oracle VM Server for SPARC, and Oracle Enterprise Manager Ops Center, as well as how to improve the performance of a single application through multi-threading.
Check it out!:
The purpose of this blog is to highlight and to explore general issues around "Cloud Computing" -- its benefits, risks, and component technologies -- and how they are evolving. I'll also periodically comment (of course!) on Oracle's Cloud Computing capabilities, resources, and cloud-related events. -- Harry J Foxwell, PhD, Principal Consultant for Cloud Computing, Oracle Public Sector HW