For a couple of months now, we’ve been trying to sort through the mess that is Microsoft’s licensing for the virtual world. Our research culminated today with a meeting with several Microsoft employees about the status of licensing and the product portfolio that Microsoft is offering. The biggest issue we’ve had is getting straight answers concerning licensing applied towards a competitor’s virtualization product. This is a complex issue to try and explain, so bear with me…
Category:
Virtualization
We are on the quest for the best thin client for our uses at HTC. My co-worker, Jason, has been leading the evaluation process. So far, different people in my group have seen presentations from Pano Logic, HP, Wyse and we have Dell coming later this month. We have evaluated the Pano Logic solution and the Wyse solution so far, and are impressed with both.
I had a fairly major revelation during VMworld. Its not something that I had stopped to think about and I guess the week of sessions and discussions helped me assemble several random thoughts together.
First, I’ve always relegated virutalization to a status where it is certainly a valuable technology and integral part of HTC’s datacenter strategy, but there is a trade-off where you have over-head and lose some performance. I realize that the trade-off is very small, but understand that our datacenter has one farm of dual-core servers – our VMware farm.
Second, I realized that our processors are just out pacing the computing needs of our current generation of software. Much of this can be blamed on Microsoft’s inability to keep their server OS up to snuff, in my opinion. But these two concepts never seemed to co-exist in my mind, until VMworld.
One keynote talked briefly about how ESX is optimized for multi-core and how it makes efficient use of gobs of memory. And, I’ve got two blade centers with blades that meet that description collecting dust as we try and figure out what to run on them. So, why at this point am I hesitant to put Exchange, SQL Server and other ‘high-impact’ applications on a virtual platform? I know that Windows Server 2003 can’t use all the memory or multi-cores. It wasn’t architected to.
We’re not making the move into Windows 2008 at this point, partially because of the Vista effect (our company is also swearing off Vista upgrades and sticking with our trusty XP). And I guess that’s a third point. Windows 2003 is just good enough for our busines. We’re comfortable with that OS, what it provides us and supporting it. I know that Windows Server 2008 has some improvements to handle the additional hardware, but I’m still hearing it doesn’t do an efficient job of it. But when you combine the complacency with Server 2003, XP and the ability to squeeze more performance out of the hardware by applying ESX, I think we have a winning combination.
In our datacenter, we run multiple instance of Microsoft SQL Server in virtual and we have very good results from this. I’m also facing a migration of Exchange onto a quad-core blade with 8Gb of RAM and thinking that’s way too much hardware for what our Exchange is doing. And that thought is a major turn – because never in my past 8 years working with Exchange have I wanted LESS hardware for a server.
And Exchange 2007’s architectual changes really has decentralized the whole Exchange “server” concept into multiple pieces which use many small servers for redundancy. That defeats the big server, multi-core, gobs of memory hardware idea that all of Microsoft’s hardware partners are pushing, except with virtualization.
So, at today’s crossroads, I’m dropping a lot of my resistence to keeping certain applications physical. And that’s a big shift in my mentality. There is too much horsepower to waste on an operating system which really can’t take advantage of it. What do you think?
One topic that I have not been able to blog about from VMworld was the VMware View announcements. I didn’t get a lot of detail, but the introduction during the keynotes did peak my interest quite a bit. My company is currently evaluating its options for doing virtualized desktops. We feel like there are some places where we can benefit from virtual desktops and others where maybe thin clients may be better. Our (possibly misguided) goal is to mitigate the amount of licensing fees we are paying to Microsoft. So at the same time, we are looking at Linux on the desktop as well.
First rule I learned at VMworld. It’s so easy to miss a lot of good sessions. Not all of my planned sessions were as stellar as some… Most were fantastic – and those are the one’s I’ve blogged about. Here is the list (with links to other folk’s blogs for the good notes) that I wish I’d attended…
- Linux Strategy & Roadmap – TA3201
- Networking I/O Virtualization – TA2644
- Virtual Center Directions – TA3807
- Managing ESX in a COS-less world – TA2659 (I was actually in this one, but had to leave because my daughter was sick & my parents were taking her to the ER)
One of the most compelling introductions I found during VMworld 2008 was the VMware AppSpeed product/plugin. AppSpeed is the market delivery of the Beehive aquisition that VMware made last year. During the Wednesday morning keynote by Dr. Stephen Herrod, CTO, all attendees were treated to a live demonstration of the AppSpeed product along with the laundry list of what it promises to be able to do.
Profile/Learn, Monitor, Mitigate
AppSpeed, in short, will passively learn your applications to map and build awareness of the dependant systems, databases (even the query level) and infrastructure services. Once it builds a profile of your application, it can then begin to monitor and watch for problems and be able to guide you directly to the root of the problem.
So, one particularly interesting session I attended was TA2275, which basically involved the speaker, a crystal ball and the future of Virtual Networking in ESX. The content was set into two parts – what we can expect in the immediate future and then some speculation about where things are probably going for the future.
One of the biggest sticking points with ESX, in my opinion, has been the VMotion compatibility and processor classes. This point has produced some rather interesting support documents both from VMware and hardware vendors. VMware has been working to alleviate this problem and demo’d the solution during VMorld again. The solution is called “Enhanced VMotion.”
The basic concept is that it takes a group of heterogenous processors, determines the lowest common denominator between the group and then sets that as the baseline for the cluster. After that point, ESX will basically mask all newer processors down to the most compatible feature set.
FT is a new feature which VMware introduced to the world during VMworld 2008. The feature is a continuous availability solution for use with some virtual machines. The following notes were compiled from session BC2621 at VMworld which introduced the forthcoming FT feature from VMware’s Application vServices. The feature should be available sometime in 2009.
FT will enable a VM to be protected with zero downtime and zero data loss due to a hardware failure. FT allows for a VM to have a secondary copy running simultaneously on a second ESX host which is executing every instruction and every input in lockstep with the primary VM. In the event of a failure, the secondary VM becomes the primary within a matter of seconds, while preserving state and without disconnecting any connections to the virtual machine. All traffic is redirected to the secondary. In addition, once the secondary assumes the role as primary, it spawns a new secondary instance on another ESX host and brings full fault tolerance back to the virtual server.
I’m back now from VMworld in Las Vegas and just wanted to let you know that the lack of posts was a reflection of the poor Internet quality in the conference location. I don’t think the Ventitian had any idea of what to expect with a geek conference. The Internet only improved on Thursday after days of lag and very slow (think dial-up) connections from the rooms and conference areas in the Sands Conference Center.
So, I’ll be working on turning my notes into posts about the different things I’ve learned during the conference and hope to have these up very soon.
