This is one of my very favorite rants. And it is one that generally has traction with people that just want to use computers, and has no traction with people that have been sucked into the way we build and define software.
I am not saying we don’t bundle up software and call that software an Operating System.
I am saying what we call an Operating System is nothing more than discription given to a bundle of software.
David Merrill of Hitachi Data Systems wishes there was a power company rebate for storage virtualization as well as for server consolidation. Link: David Merrill’s Blog » Blog Archive » A Power Company’s Reward.
Storage consolidation can save in total electric costs. Storage virtualization cannot always make that same claim. This can be attributed to virtualization encouraging the extended use of older arrays within the storage pool. Extending the useful life of IT assets is usually a good idea, but can offset plans to reduce electricity. Older arrays consume more kWatts per TB than newer systms. Virtualization with newer arrays, and providing tiers of storage (out of the frame) has also demonstrated lower kVA, BTU and kWatt consumption.
Stephen O’Grady of RedMonk has written an extended discourse on virtual appliances and the importance of the operating system to ISVs. Now I’m not an analyst, and I can’t predict the future. I do know that our initial partners in the Virtual Appliance Marketplace (mostly hardware appliance vendors) say there is demand among their customer base, and it’s easier for the vendor to control the quality of the product. It seems like a real advance in software distribution models. I think the uncertainty I’ve seen among some industry observers comes down to this: the customer makes an explicit agreement with a hardware appliance vendor to provide security updates to the OS and application, since the colored pizza box is closed. Do we trust application vendors to deliver the same level of security updates?
My conclusion, based on the above? If support matters, operating systems matter. If the subject is production systems, then, operating systems matter.
What’s equally clear, however, is that VMWare’s contention, made
here at VMWorld, that the role of the operating system is changing is
accurate. This is reflected not only in the technical innovation seen
around virtualization from both a hardware and software perspective,
but by the complexity of the support questions facing ISVs.
Unfortunately this convenience does not come without a cost. As
described in several of the sessions at VMWorld, the very notion of a
virtual appliance blurs heavily the line between application and
operating system. VMWare was, to a person, very up front about the more
limited role that operating systems play in the virtual appliance
world. Their contention is simple: general purpose operating systems
are designed to handle a number of tasks, but only some of these are
applicable in the context of a virtualized environment. Why bother
including the hardware support in your operating system layer, they
ask, if the application is being deployed to a virtualized environment?
They propose instead that application vendors create and deliver as
part of their appliance operating system layers that include only what
From an engineering perspective, this has an undeniable appeal: it’s
simplicity taken to a whole new level. What I’m wondering about is how
this will work in practice.
We’ve already cited another part of this post a few days ago, but here’s Jonathan Eunice of Illuminata on virtual appliances:
No doubt appliances have been oversold at times—never more so than
during the first Internet boom when some envisioned that hardware
appliances would replace general purpose servers for running everything
from databases to email. (Subscribers, see our The End of Cobalt and
the Appliance Era That Never Was.) And, even with infinitely “softer”
and more transportable virtual appliances, the fundamental appliance
concept implicitly assumes that appliance images don’t have to
routinely be modified, patched, or tweaked, in which case the “just
fire it up in a virtual machine” promise starts to ring a bit hollow.
Still, some types of applications (such as firewalls and web servers)
can indeed be quite standard—at least within a given enterprise—and
others have such truly ugly dependency sets that anything providing at
least some simplification is welcome. (We very much like Trac as a
project tracking system, but oh the installation process!) Thus, we
don’t view virtual appliances as the right answer for every software
deployment migraine, but they can soothe headaches for all that.
Richard Garsthagen has been burning the midnight oil to hook the VI3 SDK to Visual Basic using web services:
Well I am finding out more and more that the VirtualCenter client is by far not using all the actual functionality of the VI3 product. One of the things the SDK does expose is for every physical server what kind of CPU it has and what features are available. You can even compare if one physical server might have different cpus. So, yes I could write a very easy application in Perl (using the perl toolkit) that shows you all the servers and their cpu features, high light the differences and asking you if you want to set a specific cpu feature mask for a group of selected virtual machines. All automated. Writing that application should take me about one hour. But then all you people need to be able to run that application,
meaning you need to have perl installed, add some extra needed modules
to perl, download the VI perl toolkit, compile the toolkit and run my
small perl app. …
Yippie, I can not express how happy I am. For 3 nights
in a row I have been limiting my sleeping time on trying this
challange. But I have done it. I have a working VB application running,
where I actually understand the source code and wrote every single line
of it myself, I have no slow delay problems (as many other people have
using the WSDL file). WOW! Why did no one do this before??
Update: If you’re a Perl monger, Richard has also written about the VI3 Perl Toolkit as well.
Alessandro Perilli of firtualization.info was at VMworld 2006 and gives his wrap-up.
It’s a long and thoughtful post, so read the whole thing, as they say. His takeaways were:
- The explicit message of this year has been: software as a service (SaaS).
- The implicit message has been: focus on Enterprises.
- The overall theme has been competition.
- Among several notable sessions it’s worth to mention:
- General sessions
- Lab Manager 2.4
- Next generation ESX Server Storage architecture
- Workstation 6.0
- ACE 2.0
- Workstation for Mac OS (codename Fusion)
- What has been announced during the conference timeframe:
- A final word on the event itself
He’s very correct that the focus was on Virtual Appliances and the Changing Role of the OS, on the future of virtualization in the Enterprise, and in trying to foster growth and communication in the whole industry, including VMware’ s competitors. The most interesting to me were Alessandro’s thoughts on (1) the absence of VMware Server and
small-medium business content at VMworld — as Raghu Raghuram is quoted, we want
to work better with small business in 2007 and (2) some very practical
comments about the virtualization benchmarking work (VMmark), which
although it mirrors real-world scenarios (a mix of OSes and workloads),
doesn’t allow comparing with single-OS partitioning solutions. In the hands of an organization like SPEC, I’m sure the correct goals and outcomes can be aligned for a virtualization benchmark.
Dave Hitz, co-founder of Network Appliance, talks about two different data centers in his blog with different needs, one who has had success with VMware server consolidation, and the other who hasn’t virtualized because they are at full utilization and have good capacity management systems.
The first punchline at the end of the article is that, in fact, company #2 does have a lot of server consolidation candidates in their business, just not in their primary business application. The second punchline will come later, when company #2 realizes that VMware is about more than consolidation and provisioning — and in fact VMware is a rich platform for resource management, security, business continuity, lifecycle automation,
and enterprise desktop management.
The first customer—who asked me to describe his company as "a large UK based Telco"—used VMware to consolidate 502 windows systems onto 25 blades. He freed 173 racks worth of space. He cut power by almost 450 KW per month and reduced his power bill by $50,000 per month, not to mention reduced service and support. In all, he expects to clear 345 racks and replace them with 20 racks, with a full return on investment in less than a year.
The other customer explained why VMware won’t help him at all. He runs a large internet site with hundreds of web servers—lots of Linux and Apache. He has a load balancing methodology that lets him saturate his servers. (In his case, the systems are actually memory limited, because they cache the most commonly used data. In other environments, CPU is the limiting factor.) He argues—correctly in my view—that VMware wouldn’t help him consolidate at all, because he has no spare capacity. VMware’s management capabilities could be of some use, but he already solved those problems, so he isn’t looking for any help there.
SearchServerVirtualization.com: As we move to a highly virtualized environment, how does the role of the operating system change?
Diane Greene: The operating system shouldn’t matter to the customer anymore. What matters to them is the service they’re getting and how stable [the OS] is and how well it runs. Once you put in that virtual infrastructure, the only thing that the operating system is there for is to give the application a sort of platform to run on.
[The situation] has gotten turned on its ear where the application is in charge of picking the best operating system. Historically in this industry …Microsoft has …had a complete lock on our industry, where they control the APIs to the applications, they control the APIs to all the hardware and all the hardware devices. It is their march, people go on their timing, subject to where they decide they want to go. They are a great company and they’ve done a lot of great things, but it is to their cadence. And that changes if there’s a virtualization layer there. If all this stuff is open and interoperable, people are going to be able to mix and match, and the cadence is going to be a joint thing in the industry.
Q: What combination of factors has made virtualisation the hot topic of today? A: A lot of the excitement around virtualisation stems from problems in the current software environment – a combination of modern operating systems and applications. People weren’t happy with issues such as reliability and security. While the OS is supposed to be in charge of the hardware, people spend too much time managing them, and they’re not robust enough to run multiple applications. What was needed was a thinner layer to do the mapping onto the hardware resources.
For example, you could imagine a distributed OS that allows free flow of information between applications, or one that stops bad processes bringing down the entire OS — but we evolved not to do that.
And some of the older technology was too slow — now the hardware has arrived.
Sun’s Warren Ponder was at VMworld and had so much traffic his voice gave out by day 2:
The big buzz was unquestionably desktop virtualization. Diane Greene – President of VMware made it very clear in Tuesdays keynote that desktop virtualization is happening now and taking off. I strongly recommend you visit VMware’s website and look at the presentation from the keynote.
The feedback from existing Sun customers and the ones that had no idea we offered compelling desktop solutions, was very clear. WOW you guys are doing some really cool stuff. Of course, I am slightly biased and agree we are doing some cool stuff. We also have some work to do too make it better in my opinion. There was only a small handful of people that had requirements that we could not meet and I actually recommended they consider another solution. We are not going to be the best fit for everyone and this is OK. I did find it ironic though, one of them kept coming back with more questions.
I have a pretty good idea of what others are doing with their solutions. Most of which, I have started calling a "Desktop on a Desktop". Guess what? The customers are as well! Customers are smart, they are not stupid and I knew it was only a matter of time before they bushwhacked through most of the hype around some of the other solutions out there. Several people made it very clear our solution is the only solution that provides a seemless user experience and appears to be the most streamlined. I try to be fair, we know where our gaps are and I really make an effort not to hide them. I make it clear where we feel we need to improve and listen closely to customers about where they would like to see our solution improve.