In her Chronicle of a Feature Foretold, posted over the weekend, Lori MacVittie gracefully alerted us to fact that both her company F5 and Cisco were planning big announcements at this year's VMworld so naturally I've been keeping tabs on this while putting the finishing touches to the marketing campaign for our own announcement scheduled for September 14 at High Performance on Wall Street (Twitter #hponws.)
What is fascinating is that both announcements address the same fundamental challenge - application mobility - in broadly the same way. Both reference cloud as one of the key drivers which is absolutely fine, but then present what to my mind are complex solutions.
First there was F5 with their Live Application Mobility press release yesterday, which Lori covered in her official capacity in Migrate a live application across clouds with no downtime? Sure, no problem. When I say complex what I mean is the sheer number of components needed to make this magic possible. The press release reads like an Oscar acceptance speech -
An agile Application Delivery Network is the fundamental component required to extend virtualization beyond a single data center or VMware vSphere 4 instance. This integration uses F5's BIG-IP® solutions with VMware vSphere 4, VMware VMotion™, VMware Storage VMotion, and the VMware vCenter™ APIs to overcome many of the networking hurdles that previously prevented organizations from migrating live applications to and from the cloud. The integration between VMware vCenter Server, BIG-IP Global Traffic Manager™, and BIG-IP Local Traffic Manager™ simplifies global traffic management to enable easier live migrations. The solution also takes advantage of BIG-IP WAN application delivery services to accelerate VMware VMotion and VMware Storage VMotion events over links that are limited by bandwidth, latency, or packet-loss.
There is nothing wrong with acknowledging the role played by all these components but I think you'll agree that there are a lot of them and the solution is heavily caveated in Lori's blog post.
This announcement was closely followed by one from Cisco on Long Distance VMotion published late last night PST. Not to be outdone this post comes complete with a link to a white paper. I rather like Omar Sultan's comment -
Looking at the jointly validated architecture below, one of the cooler things to point out is that, for Cisco customers, at least, the solution builds upon the gear they probably already have in place.
The white paper does a thorough job providing the details of the networking chicanery that is now possible which unfortunately, combined with the earlier press release from F5, brought on an epic migraine.
Just kidding. Nevertheless reading these posts I was struck by the thought that both F5 and Cisco illustrate perfectly Maslow's Maxim -
If the only tool that you have is a hammer, you will see every problem as a nail.
In other words ever since VMotion was first announced way back in 2005 the holy grail has been to extend this to solve the general problem of application mobility.
If ever there was an example of the wrong level of abstraction being applied to a problem this is it which is why four years on only limited solutions are being announced. It is also why the Cloud Interop discussion tends to get bogged down in discussions concerning VM formats etc.
The solution to this problem? I would have thought this was clear but essentially it is about focusing on the middleware and delivering this in cloud-speak bundled as platform-as-a-service offering. Essential dealing with the cloud interop issue by making the problem go away (what Wittgenstein once called untying the knot). That's where we're headed as a company as that's what our announcement will be all about in a couple of weeks time when we get together with our partners in New York.
Recent Comments