Subscribe

The end...

...of infrastructure as we know it.

Samantha Perry
By Samantha Perry, co-founder of WomeninTechZA
Johannesburg, 23 Mar 2010

Infrastructure as we know it has changed forever. CIOs today have to manage infrastructure that is not theirs, that is not on their premises, that belongs to companies they've never heard of, or doesn't even exist, strictly speaking, in the physical sense of the word.

Cloud computing, and cloud services, like Google Apps, have arrived, and are changing IT completely. Meanwhile, end-users don't really care that it's Google's service that's slow; they expect you to fix it. And so does your CEO.

There is light at the end of the tunnel, however, and it's not the demise of the cloud. Says BCX Services GM Julian Liebenberg: “Today's infrastructure can no longer be defined in terms of technology and geography. With more and more clients adopting cloud computing, they are beginning to realise that the infrastructure cannot be owned, touched, or even contracted to deliver a required service level.

“Yet we are able to manage this seemingly ubiquitous infrastructure in such a way that it supports business processes in a controlled and reliable manner. End-to-end management is now achieved by managing the delivery of the business process from its point of origin up to its point of consumption, at the same time adding services such as bill presentation and security, and ultimately delivering a complete service.”

Which is all fine and dandy if you have an outsourcer to whom you can hand over the problem.

If you don't, says Todd Schoeman, BT Global Services head of MEA Portfolio & Marketing, you'll still be entrusting your cloud services to a provider who will be bound by SLAs (service level arguments, or agreements, depending on which company you outsource to).

It's all going financial.

Todd Schoeman, head of MEA Portfolio & Marketing, BT Global Services

As Liebenberg puts it: “When customers use modern infrastructure owned by a multitude of vendors over which they have no control, and comprised of the complex technology architecture needed to support business processes, management that surpasses the limitations of protocols, SLAs and geography is crucial.” That's easier said than done, of course.

That it's complex is something of an understatement. Continues Liebenberg: “In the traditional outsource market, it was always 'fine old-fashioned' infrastructure management. Either you owned it or knew who owned it, so you could define an architecture and what service level it needed to provide. Managing it was as simple as selecting a tool, and getting into an agreement with the owner to gain access to the infrastructure so you could monitor and report on it. With cloud you don't even know who it belongs to, so it becomes challenging to monitor activities on that.”

Moving mountains

Although complex, says EOH director Rob Godlonton, IT has become a service-based commodity.

“Business just wants it to work,” he says. In four or five years, business will expect IT infrastructure to work the way the lights do, and it will expect to pay on a service base, he adds. The infrastructure being rolled out now, both data centres and broadband, will allow clients to obtain infrastructure on a hardware-as-a-service, software-as-a-service, and applications-as-a-service basis, he continues.

“The challenge for the CIO is to decide what he wants. And organisations like us need to step into the information gap and put together service offerings that leave the CIO free to worry about what makes the business really tick.”

“You don't have to manage the infrastructure,” says Schoeman, “you manage the SLA. The infrastructure doesn't really exist in a physical sense, so you need to entrust the SLA to a partner. This is usually an outsource partner, but it can be a single source or multi-source arrangement. There are benefits to both,” he notes. “Single source means one SLA and one butt to kick, and one body to negotiate with if things get bad. If you go multi-source and isolate services into different towers, eg, use a best-of-breed networks provider, and applications providers and so on, you need to decide what's core or not. And do you want the competency to manage that inside or outsource it? In which case you'll be going to one provider (called an aggregator) anyway.”

Moving away from traditional thinking around ownership has been hampering adoption of cloud services, and will continue to do so, but, says Liebenberg: “Once clients understand you don't need to own the infrastructure to have a managed infrastructure, it opens up a whole new ballgame. We believe you don't need to own infrastructure to have an SLA on it. We're offering SLAs on the Internet, on the cloud.

The challenge for the CIO is to decide what he wants.

Rob Godlonton, director, EOH

“We've been talking for some time about managed service providers, where the Internet is one such provider. Now customers have access to many more, for example, a local ISP for Internet access and Google for e-mail. How do you provide service to that user? The user can't phone to ask, 'How do I do this on Google?' It brings a new flavour of multi-sourcing where some of the vendors are known and others unknown. It's more challenging to provide than a full outsource, where every aspect of the infrastructure has some accountability to someone.”

More challenging it is, but it's where the entire world is going, with the exception of institutions like banks, which have obligations in terms of keeping customer data within defined geographical (or geo-political) locations, and even they will be investing in private clouds if the vendors are to be believed.

Will everything go cloud? Says Schoeman: “It's all going financial. The bean counters don't want to depreciate capital expenditure (capex) off financial statements, and the minute you buy and own infrastructure, you have to depreciate capex, which drops your earnings before interest, tax and depreciation (EBITDA), which they don't want. So the first benefit is that the provider gets the capex bill and you rent it as a service. The books look better, EBITDA is better, shareholders are happy. The second reason is that companies want to stick to what's core, and outsource non-core activities. If an IT service is not part of a company's core function, then they will get rid of the IT expenditure as an operating expense (opex). Providers get economies of scale out of servicing multiple customers on shared infrastructure, and delivering services out of the cloud; end-user organisations don't.”

Looks like the accountants win it again...

Share