RTFM Education – Virtualization, VMware, Citrix

Jun 30 2011   12:17PM GMT

Tech Field Day – June, Boston

Michelle Laverick Michelle Laverick Profile: Michelle Laverick

Well, the dust has settled – or rather I’ve been so max’d out that I haven’t had time to write up my experiences. And I really should before it gets all fuzzy, like afternoon beer…

This TFD was in Boston, and virtualization focused. I was joined by a group other Merry Pranksters.

I found the event immensely fun, and tiring – and I wasn’t partying down at the bar with the other guys at the end of each day – and I was grateful to Clare & Stephen for arranging the events and making sure got enough R&R at the end of each day – be ready for the next…

The quality and range of presentation was many and varied – and think some vendors were a little intimidated by the group. TFD has a reputation for not suffering fools gladly, but I think occasionally our bark is worse than our bite. The presenters who came off worst were the ones who hadn’t really found about out the delegates – they came across badly when faced with Edward Haletky who’s a bit of security whizz – another vendor started to talk about VMware SRM not realising that I was in audience. Anyway, I don’t take these things personally (in a ‘don’t you know who am I way’) but no-one likes it when a presenter assumes you know nothing about subject – because they didn’t bother to find out the audience profile. TFD delegates aren’t pro-analysts or journalist (even though I find myself being some sort journo right in my current gig) – so delegates don’t especially warm to the kind of “briefing” content that’s some companies rely upon…

The general theme that came out from the event – is that group wanted more focus on integration with VMware vCenter – from all the vendors. Not a session went by when some didn’t mention whether the software in question had plug-ins into vCenter. That was especially true of the performance vendors – where folks were concerned about bringing in tools where VMware Admins just might not crank up the software – so overwhelmed they are by additional monitoring systems…

Personally, I would have liked have seen more small start-up vendors who were doing real bleeding edge development work. Perhaps that could be the a theme the TFD could pick up on next – Start-up TFD…

Anyway, here’s my decidedly partial view of each of the vendors who presented to us…

I was very impressed with vKernel’s presentation. The young guy they had up was a very strong presenter, who was backed by the senior management team. I’ve been meaning to have the vKernel guys on the vendorwag for sometime, and for reasons of tardiness which is entirely my fault that didn’t happen. Anyway, we’ve reconnected, and did another webex, and the questions are submitted – so you should hear from them soon. I was impressed with the maturity of their offering. I was particularly impressed by their ability for folks reserve capacity for future VMs that will be deployed at some later stage. Effectively, masking those resources from the rest of the system from a capacity planning perspective. I thought that was really quite smart. Of all the virtualization performance folks we saw – they seemed to have the most rounded offering. I’m not a fan of having a performance tool for each every resource – a disk perftool, a network perftool, and so on. I know the customers who I speak to will struggle to raise a budget for each of these tools – but they could perhaps raise the money for one tool that did the whole kit and caboodle…

Were treated to three different presenters at VMware covering three different technologies – the VMware Mobile Virtualization Platform, Site Recovery Manager and Student Cloud.

By far the most intriguing and engaging of these was on MVP. It was first demo’d in 2009, and since then there’s been little said. Here we had working demo of Android based phone, with a type2 hypervisor embedded into it – running a second copy of Android. The idea is that you can have one device that is both you “home” phone and your “work” phone, and the corporate space can push out updates and secure the “work” side. There was some debate about whether there is a market for such a development – but as folks increasing want to bring own favourite device to work. Just speak to the UK VMware guys who are forced to have BlackBerrries, there’s talk of them getting a Stanley knife blade on the SIM card, to have fit inside their iPhones!!!  In similar way  Android on Android is great, but doesn’t really help the folks like me with an iPhone. I wouldn’t beat VMware to much on this – after all the iPhone isn’t “open”, and can’t be run on anything other than Apple device. Until that changes – its hard to see it getting mass appeal until when (not if…) Android supplants the iPhone as the most popular OS on a smartphone. With that said, Apple does allow Apple OS X Server to run under VMware Fusion. So who knows there might be movement on that.  I would have liked to see less of the “platform” and more about the management side of things, which unfortunately we didn’t get a chance to see.

I was very disappointed with SRM presentation – which seemed to replay features, functionality and usage cases that would have been more than familiar to my fellow delegates. There’s was just one PPT on the new “vSphere Replication” technology which I’ve written about and spoken about at VMUG events around the US this year. I saw this as missed opportunity to really stoke up my fellow delegates about some of the new features – if you come along to my presentation in July in Indianapolis you will hear more about what’s new in SRM in 10 minutes, than we did in half-an-hour.

Finally, VMware showed us their “Student Cloud”, a revision of vCloud Director – geared up for University, Colleges and educational establishments who need to spin up VMs which contain “lessons” for student. I wasn’t really clear who those students would be – information systems students I guess. I’m assuming the proximity of MIT to VMware’s MA offices inspired this fling which will be on the “VMware Labs” site. I’m not quite sure that there’s a mass market for the technology, and there’s wasn’t a great deal of clarity around licensing – but one would assume that educational establishments would qualify for the same sort of discounts that they currently receive from VMware, should they chose to adopt it.

Solarwinds recently acquired Hyper9 – so we had a both a Solarwinds and Hyper9 representative talking to us. I really felt for the representatives who seemed to be at distance from my fellow delegates – and stayed locked behind a desk & laptop doing the PPT/Demo piece. Twice they were asked to come closer, so folks who couldn’t hear them because of the aircon – could hear what they had to say. Things might have worked out better if one of the guys had “driven” the demo, whilst the other stood and spoke – and vice-versa. But I’m guess each guy was a specialist in their product and that would have been difficult to finesse.

Symantec talked about two products – ApplicationHA and BackupExec. I was really very interested in ApplicationHA more than BackupEXEC. I don’t have single customer who uses BackupEXEC, if they do use Symantec its NetBackup they are using. I’ve used BackupEXEC in some of my early books as example of a 3rd partner to the now defunct VMware VCB. But that’s about it – to be honest I didn’t see much in the way of innovation on the backup side – more catch-up with what companies like Quest (nee VizionCore) and Veeam have been doing for some time. I guess that’s to be expected – when you have a company that backs up EVERYTHING, compared to a company that backups just VMs…

Anyway, ApplicationHA is where it was at for me. Basically, they are hooking into the new API’s provided by VMware HA in vSphere4.1 that allows for 3rd parties to inspect the guest operating system for service availability – to then take remediation action – e.g restart the service. The integration is very slick and smartly done. OK, Yes, basically all they are doing is monitoring the service, and giving it a kick when/if it fails. And you could say any good script-monkey could write something like this – but the point is this has all been done for you – through a UI which is directly integrated into the vCenter [There’s that vCenter integration thing again!]. I’m hoping to play around with ApplicationHA sometime in the future – right now I can’t because the .ahem version of vSphere I’m running is currently incompatible with it. That’s sure to change in the VERY near future…

What’ I’m not sure about is whether customers would be satisfied with ApplicationHA (or it nearer neighbour from NeverFail). Wouldn’t customers prefer the full fat of Veritas Cluster Server, MSCS or NeverFail – where there is two nodes, and failover occurs between one node another?  I guess the ultimate measure is who delivers the best up-time, at competitive price – with low complexity.

The reason I’m really interested in this stuff. Is I know SRM is NOT an availability tool. It’s a DR/DataCenter Migration tool. I know customers will look for availability tools deliver up-times to their services. That’s territory that currently companies like VMware shy away from. But if it really is all about the application and service – its hard to avoid a situation where state of services is not being monitored and managed. Of course some might say the future lies in “cloud aware” applications/services which have built-in availability that leaves conventional clustering redundant. That’s a great idea, and I’m such a creature exists, but until a Unicorn trots up to my paddock – I think we might be using such in-guest availability tools like this for sometime….

Zerto were our secret/mystery company – who at the time were still in “stealth mode”. [That’s a new phrase on me – but since TFD I’ve heard it used 10-20 times…] to describe how they weren’t public. Zerto is pun on Zero RTO. Geddit?

I was impressed with Zerto, and given my interest in VMware SRM – I’m keen to see how this market segment will develop over time. For example I’ve spoken at length with the guys from virtualsharp.com. The focus was very much on their new replication appliance which is very like VMware’s vSphere Replication. I was particularly interested in their capacity to monitor and throttle bandwidth used by replication. The focus was very much on this appliance  – which uses VMware’s CBT as way of tracking changes to be replicated. VMware’s VR doesn’t use this currently – so that was an interesting deviation. Less was said about the process of recovery – recovery plans, failback, re-IP, script call-outs etc. Anyway, I’m hoping that once the burden of writing the new SRM book is through – I can have the Zerto guys on the vendorwag to find out more. After all replication is easy, automation is hard…

Embotics style themselves as private cloud management player. Personally, I hadn’t come across them before. They started well, but then seemed to falter under questioning from my fellow delegates. I found it tricky not to just seem them as occupying the just-another-portal territory. I’m sure that’s probably unfair. The positioned themselves as rapid deployment to the private cloud – something that makes me feel nervous. The two phrases “rapid deployment” and “private cloud” should not be seen too closely together. Otherwise the project can become a CEO tick-box activity. Cloud. Tick!

I think the important thing about any cloud automation layer is two things – it must bring a genuine separation of the self-service portal from the virtualization layer – something that offers a clear abstraction of virtualization constructs (the cluster/host/resource pool) and presents to the cloud layer brand new set of objects with which to set resource allocations. Without these second tier of virtualization what you might get is an second unpleasant  thing – just another layer of complexity.

I don’t know – this year I’ve spoken to a lot of “cloud automation” vendors – and they ALL have the same narrative. So perhaps I’m getting bit jaded. If you heard one cloud automation pitch, you’ve heard them all. There’s very little deviation from the script – or innovation. That’s what’s makes me interested in Nimbula who are currently using KVM as their hypervisor (more to follow), and have their own method of virtualizing networks without being dependent on VLANs alone… They should be on the vendorwag soon…

Finally, we saw NetApp. In reality we spent more time with the folks from Akorri who NetApp acquired recently. I actually quite liked the Akorri BalancePoint product although I found its UI somewhat dense. I guess that’s UI familiarity for you. What I really liked was the way they could show visually the path from the application (say MS – SQL) thru the virtual disk, the hosts HBA, the switch, the Controller, Array and Volume. I don’t think I’ve seen that kind of visibility at the storage layer before. However, there was decidedly less info on the CPU/Memory/Network front. So as performance analyzer all rounder it was less convincing. Of course with further work and development – I’m sure they will “get there”, the trouble is that some other vendors are “already there” now… My only disappointment is that we didn’t also see EMC in Boston as well as NetApp, as I have good relations with both organizations.

 Comment on this Post

 
There was an error processing your information. Please try again later.
Thanks. We'll let you know when a new response is added.
Send me notifications when other members comment.

Forgot Password

No problem! Submit your e-mail address below. We'll send you an e-mail containing your password.

Your password has been sent to:

Share this item with your network: