IBM gets smart with cloud

Opinion Clive Longbottom Jul 29, 2011

IBM's SmartCloud is examined by a leading analyst. His conclusion: if IBM can keep it simple, it could be a winner

As one of a rush of vendors that have been jockeying for position in the race towards being a major cloud vendor, IBM has made another set of announcements over its own approach to cloud computing.  It is being done under the banner of SmartCloud (part of IBM’s overall umbrella message of “Smart” – Smarter Planet, Smarter Cities, Smarter Computing, with more to come).

Based on an IBM Cloud Computing Reference Architecture combined with an IBM Common Cloud Management Platform, IBM’s approach is to provide a full multi-tenant stack with the flexibility that IBM and its partners need to meet the specific needs of the end user client.  On paper, the reference architecture looks exceedingly complex, but this is not a portfolio sale.  IBM is positioning SmartCloud as an infrastructure as a service (IaaS) or a platform as a service (PaaS) play, and although the complexity is there within SmartCloud, it should be hidden from and irrelevant to the end user, who will be utilising SmartCloud through a self-service front end.

IBM acknowledges that, counter to some vendors’ messages, this is not an argument between private (only used by the one organisation) and public (available to all) or shared (available to a smaller group) clouds, but is a discussion that covers a spectrum of different options from private cloud only to public cloud only, with the majority of users ending up with a hybrid usage model somewhere in the middle, with some workloads being served by internal systems, some from commercial cloud providers and some from public, “free” cloud service providers. SmartCloud provides the shared cloud part of this spectrum, but IBM is more than likely to take the reference platform and use that as the starting point for working with its larger customers when working on the provision of private cloud systems. Indeed, IBM posits five different cloud provision models, as such:

  • Private cloud, owned and operated by the customer
  • Private cloud, owned by the customer, but operated by IBM (or another provider)
  • Private cloud, owned and operated by IBM (or another provider)
  • Virtual private cloud services (based on multi-tenanted support for individual enterprises)
  • Public cloud services (based on the provision of functions to individuals)

SmartCloud comes in two different flavours:  SmartCloud Enterprise aimed at providing a mass-market, shared environment based on speed to capability, with the resources and pricing primarily being based on usage, and SmartCloud Enterprise+, which offers more resilience, including guarantees of 99.9 percent, and the system is fully managed with multiple layers of security across the virtual and physical systems

 

SmartCloud Enterprise offers basic virtualised cloud platforms based on Windows or Linux, with availability promises of 99.5 percent.  Basic virtual machines complete with operating system are provisioned on a pay as you go basis, with the capability to load up additional software – on either a “bring your own licence” or a pay as you go licence model.  The big issue that Quocirca sees here is that there appears to be some inconsistency across IBM’s information sources – and that this could lead to confusion within the direct and indirect channel, as well as with the prospect base, unless addressed rapidly by IBM.  Security is provided in depth at the virtual layer, with a good degree of physical security.  Management is essentially self-service with premium support available at extra cost.  Although SmartCloud Enterprise is termed by IBM as IaaS, Quocirca would say that it is a functional PaaS model, as the operating system is pre-provided. The platform is aimed more at the development and test communities requiring a fast means of trying something out at a cost that can be highly granular, being charged on an hourly basis. However, batch analytics is another area where IBM is seeing some traction, due to the predictable costs of an hourly usage model.

With SmartCloud Enterprise+ users are offered  AIX in operating system mix, enabling complex workloads to be supported. Pricing is based on monthly usage-based payments over a fixed contract period, and the system is aimed far more at run-time environments. Still hosted from an IBM managed data centre, customers can choose either a dedicated or shared system (based on an IBM “pod”), dependent on their risk profile and needs.

As both systems are based on the same underlying reference model, it is easy to move workloads from the SmartCloud Enterprise to the Enterprise+ platforms, making the move through development to testing to run time relatively painless.

In both Enterprise and Enterprise+, users can choose a data centre in a specific location should they need to meet geographic specific legal requirements.

So, just how successful is SmartCloud likely to be?  As an IBM-branded offering, it is unlikely to resonate immediately with the mid-market and small and medium business (SMB) buyer.  The “Big Blue” perception is still too entrenched, and IBM will need to bring new channels to bear, enabling SmartCloud to be branded by the channel to have much chance of penetrating this market. IBM promises that such “white labelling” capability is on the road map.

Large enterprises are a different issue.  For those with deep existing investments in IBM, SmartCloud can make a great deal of sense.  However, IBM may want to look at how it can bring its IBM Global Business Services (IBM GBS) and IBM Global Technology Services (IBM GTS) teams into these customers to help them migrate their existing IT estates to the IBM cloud reference model, so that workloads can be moved far more easily between internal and external environments as required. In conjunction with IBM GBS’ existing business transformation capabilities, SmartCloud can offer a flexible, scalable and enterprise class platform – for those who can afford such an overall approach.

As an enterprise response to the likes of Amazon EC2, SmartCloud looks promising. IBM’s biggest issues in the past have been when it brings something relatively simple to market and then rapidly introduces complexity.  It has to accept that for SmartCloud to be successful, it has to keep it as simple as possible.  If it wants to bring a new service to market, then this has to be packaged completely and offered as a specific business value option.  For example, its IBM SmartCloud for SAP applications packages SAP as a service within SmartCloud – and keeps it all relatively simple. 

If IBM can maintain this approach, and not start to offer, let us say, IBM SmartCloud Collaboration Edition (with added Lotus bits), IBM SmartCloud Analytics Edition (with added Cognos goodies) and so on as defined and separate SmartCloud offerings, where there is a definite need for such systems to integrate and operate with each other, then SmartCloud should be successful – in large organisations.  At the lower level, IBM will need to think long and hard as to how it will allow SmartCloud to be licensed, operated and sold through the channel.

Clive Longbottom

longbottom.jpg

Clive Longbottom is founder of Quocirca and a globally recognised industry analyst.  His primary coverage area is business processes facilitation, covering the core processes in company’s value chains.>

Within this remit, Clive covers a range of approaches and technologies, including collaborative tools, workflow, service-based architectures, data centre design and optimisation, hardware and software stacks and outsourcing, including cloud computing.

Clive is a regular presenter on public platforms and is consistently placed high in Apollo’s “top European quoted analysts” tables.  In his spare time, Clive is a fully qualified wolf handler, working as a volunteer with the UK Wolf Conservation Trust based in Beenham, near Reading.