Managing the Cloud

23 July 2014


I get a strong sense that CIO’s across New Zealand want to make changes but feel burdened by their existing provider and service constructs. Ironically the situation is the same whether they do the work in-house or outsource it to a trusted provider.

CIO’s and Service Managers are also expressing increasing levels of concern around the Service Management and Service Desk environment needed to support private, public and hybrid cloud instances. Many providers seem to be struggling to monitor and support an infrastructure and services environment that has been designed by someone else and is located outside their control in the Public Cloud.

This difficulty is not unique to New Zealand. I have also had a series of conversations with some very large organisations struggling with the same issue in the US and Australia. So what is the cause of all this angst?

Cloud service management is not the same

  • New Zealand Service Management and Service Desk environments have often been designed and refined by providers who associate control of the managed assets with locked in income streams of their own. These can be in the form of fees or salaries.
  • These systems and processes operate via enterprise management software (ESM). ESM products are extremely expensive, painful and time-consuming to design and implement. They leave a legacy of cost that has to be recovered somewhere from someone.
  • Customers have often invested large sums in transition to these provider-delivered ESM’s, so they also have money and sweat tied up in the service. Change can be expensive and politically tricky.
  • Given the above, there is an understandable fear felt by both provider and client of having to reinvest in some new, as yet untested, cloud service management construct. The result: both provider and client have an anchor limiting their own and their business’s trajectory and speed.
  • These services and processes appear to be tied at every level to the Enterprise System Management platform and are then aligned with the complex ITIL methodology. This is often over-complex and assumes complete, end to end control by a single partner. This is unhelpful when a client wants to access cloud based SaaS, PaaS and IaaS from a selection of providers from around the world and locally. (A hybrid model)
  • The existing ‘control and command’ service is highly desirable in the technology sense, especially for the provider. However, this model risks numerous service management collision points and additional cost to all parties while needlessly repeating the service management work already done behind the firewall in the cloud.

Navigating a path to cloud service management – key points

  • It’s new. The challenge of provisioning and managing a third party, off-shore, public cloud environment requires different thinking. The services and delivery constructs which are needed are not as clearly defined or financially transparent as they need to be.
  • It’s new. Even identifying where the demarcation lines for Service Management among the incumbent provider, new third party providers and the client has been a challenge to date. Unless we understand where responsibilities begin and end we cannot achieve a cohesive service management model.
  • It’s different. Overlaying the existing local Service Management model with a public cloud model implies that it will cost more to run the service overall. (1+1=2). This is both illogical and financially untenable. We are, after all, already paying by subscription for the cloud provider to design, build, deliver and run the cloud-based infrastructure, platform or application. So why pay twice?
  • The risk is proliferation. Service Management costs could and should in my view be fixed or near fixed costs when dealing with cloud based SaaS, PaaS and IaaS. If this were not the case the provider will be forever incentivised to allow instances to remain turned on months or years after they were no longer required.
  • Those of you with ‘magpies’ working in IT who, for reasons of false economics, refuse to switch off existing servers until they die, will be familiar with the expression ‘we will end up with data centres the size of bowling alleys at this rate’.
  • Virtualisation offered the promise of fewer servers through consolidation. However the reality is that most virtualised IT environments have many times more instances than they started with – to little effect. This only benefits the purveyors of virtualisation software, operating systems and service and support contracts. We run much the same proliferation risk with cloud based server and storage instances.

My next blog will include some suggestions as to how to address some of these challenges and obstacles. If you have any thoughts/feedback, feel free to comment.

Read the next blog in this series:

Cloud Services Management Part Two

Read the previous blog in this series:

Cloud Service Management