Clear Skies for Data Center Virtualization?

I just finished reading the latest Strategic News Service (paid newsletter) about cloud computing and it took me back to a recent Interop conversation.  While Mark Anderson was talking about a massive shift to cloud computing and its impact on the server industry; I was also thinking about Alistair Croll and Blue Lane’s Allwyn Sequeira getting rather animated about what impact cloud computing would have on security and traffic management; and I knew it wasn’t because of the Pinot.

 

When I blogged about virtualization-lite, or the proliferation of hypervisor VLANs, the notion of constraints and limitations (and their impact on markets) really struck me.  If Mark Anderson is right about the computing world becoming server farm-centric, then what impact will that transition have on the network and security hardware business? 

 

I think it’s likely that cloud computing would bring back the strategic IT era in the form of new and innovative service providers.

 

The recent Amoroso (AT&T (T) CSO) interview offers a glimpse.  One piece of the cloud would obviously be security as a service.  It would have to be.  Imagine the impact of a security breach in a highly fluid hypervisor mesh in one of these super-sized data centers?  Service providers who can monetize security would be less likely to suffer through the risks and watered-down rewards of virtualization-lite or be slaves to compliance table stakes initiatives. 

 

Network security as a service could be better than the decentralized model where most enterprise security postures accept significant risks and are only aimed at being slightly better than another’s.  This is a similar dynamic to the insecure software game.

 

Yet a hypervisor backplane would also be a mystery to most application front ends and load balancers.  Without visibility into the hypervisor layer their ability to manage traffic and policies between VMs would have a similar disadvantage as the network security players blinded by hypervisor rack and stack.  So is traffic management the next chokepoint for the virtualization of the data center?  Could it be another driver of virtualization-lite?

 

If service providers can monetize traffic management they could also become strategic in this area.  They would be encouraged to drive vendors to heightened API access and enhanced application layer traffic management functionality and away from virtualization-lite’s hypervisor VLANs.  They would also have enough leverage over vendors to drive strategic innovation, unlike our current scenario of scattered midmarket enterprises arm-wrestling for existing features from the previous year’s slide ware.

 

The virtualization of the service provider server farm may be the trigger for true data center virtualization (including VMotion) because service providers would have enough clout and enough business case to make their vendors innovate; to make virtualization in production environments really work as promised.  In addition to the Verizons (VZ) and AT&Ts, such a market might also include Google (GOOG) and Amazon (AMZN) and others.

 

The enigmatic challenges of virtualization security and virtual traffic management would be solved.  Then more innovative solutions would ripple through the enterprise either as a service or as a truly new virtualized data center.  If Mark Anderson is right, that would be the equivalent of a grand slam for a virtualization platform vendor, its service provider customers and the security and traffic management vendors who position themselves to land the business.

 

It also could have a dramatic and sustainable impact on global power consumption, network and Internet security and IT flexibility.  This is sounding more and more like an election issue…

 

Note: You can find out more about the Strategic News Service here.

Disclosure: none