Tech direction has become very difficult to predict recently - cloud seems all pervasive, but how many companies have successfully managed to implement cloud in a way that will make everyone (users, management, audit, regulators, security etc.) truly comfortable?
Has the CIA's contract with Amazon finally sealed the deal?
Will concerns over data soverignty mean the mega data centres of some cloud providers fall short of requirements and see the emergence of medium sized data centres in multiple locations?
Will this mean that cloud providers actually share the same facilities to still get some scale?
Will the business case stack up, when all the depreciating assets that are still on the books are added onto cloud service costs?
Will IT management (and risk management) ever see their required level of control over the shadow credit card IT economy in their companies?
Will those who have had services hosted continue to call this cloud?
How will companies manage the multi-sourcing environments they're (in some cases sleep-) walking into?
What happens to your customer service when a cloud provider goes out of business?
What happens to those companies who have based their business on hardware implementations?
Why is VMware not prevalent in cloud provider stacks?
Is OpenStack, CloudStack or Microsoft going to prevail?
Who will hardware providers sell to? End user companies? Software service providers? Infrastructure service providers? Will they attempt to set up their own "cloud" service offerings? Which of these hardware vendors will prevail, which will flounder?
How quickly will legacy proprietary systems fade away?
Over the previous 2 to 3 years, VMware seem to have been lacking in direction, raising probably more questions than were answered as they purchased innovative products, but failed to integrated them convincingly into a clear strategy, nor into their existing product set. Even venturing into app dev platforms, although this is now spun off into a separate joint venture. Much of the direction taken by VMware announced over this year's VMWORLD conferences has some clarity and distinctly more focus. Their more recent acquisitions are fitting into their direction and have been (at least on the face of it) integrated into their products. Throughout the conference there was a consistent message - everyone had clearly been briefed.
The key disruptive message for me were:
- NSX - virtualising the network. This is a centrally managed but server distributed implementation which layers over the top of physical networks. This is fairly basic in its capability at the moment, but will add features over time, and bcause those features are in software, it will run over commodity switching hardware. The hardware will still need to be reliable and performant, but it won't need lots of processing capability to rund a proprietary software defined network protocol. For smaller IT shops who are fully virtualized and for service providers who truly need fast agility then this is a great opportunity. For those locked into proprietary switching hardware and a fear of the new, this will take some time. It does look compelling though, as long as it does what is says on the tin. Good to see continuing support for non-VMware hypervisors too.
- vSAN - virtualizing the capability of a SAN across the internal disks in your VMware hosts is not a unique or new ideal (Nexenta, ScaleIO for example), but having a tier 1 vendor version will mean it will take off rapidly. I can imagine the server admins out there being enthusiastic about breaking free from their SAN and storage team so they can implement faster than before. Costs will fall too. We'll have to see how quickly the HCL builds, but it should be pretty quickly. Initially workloads will be limited to those that have a fairly predictable and fixed ratio between storage and compute (e.g. VDI), but I'm sure flexibility will build over time.
- VMware Hybrid Cloud Services - vCD didn't take off in the service provider community. VMware didn't win that one, so they've joined in the fray. In theory this should make it easier for customers who have their internal systems running on VMware to choose VMware as their hybrid cloud provider. In theory. Key here will be just how VMware licence the tools required to integrate internal and external clouds. If they make the customer's internal costs to implement (remember, there won't be that many customers with vCD / vCAC / ITBM in place internally already), then this project will struggle. If they get it right, then will Amazon, Microsoft and VMware be the main players of the future?
So many possibilities, so much choice, so much opportunity - to get it right, or to get it very wrong!
No comments:
Post a Comment