Managing multiple clouds: how to make the process easier

Advice Rene Millman Feb 1, 2013

With many organisations opting for multiple clouds, how do CIOs ensure they have things under control?

As the cloud computing market matures, more organisations are finding that one cloud does not fit all. Indeed, as with traditional infrastructure, CIOs may well make a strategic decision not to put all their faith in one provider. There’s also the ‘shadow IT’ factor, the phenomenon where the heads of lines are choosing cloud services, sometimes with the co-operation of the IT department, but often not.

For whatever reason, CIOs will often find themselves managing multiple clouds. How should IT professionals ensure they can administrate and oversee all the different relationships that come with multiple cloud providers? In particular, how can organisations cope with the complexities of dealing with public, private and hybrid clouds.

From the outset, managing any cloud requires the organisation to have in place sensible service level agreements. Being clear on the level of service you expect to receive and managing that effectively is important, particularly when multiple service level agreements (SLAs) are involved, according to Colin Bannister, chief technology officer at CA Technologies UK and Ireland.

“A focal point for many businesses in recent months is building on SLAs to ensure they cover every eventuality which could impact the service you receive; from natural disaster to human error,” says Bannister.

“Experience tells us that we need to stipulate requirements under all possible circumstances which could affect the smooth running of your business.”

How should IT professionals ensure they can administrate and oversee all the different relationships that come with multiple cloud providers

For example, what impact would five hours downtime have on your business and how will your service providers compensate any losses incurred as a result? Bannister says  managing multiple cloud providers will certainly require clear agreements and solid management across the board.

“In an ideal world you will have one overarching agreement in place that all providers must adhere to. Simply because when you combine the nuances of several you will soon find that management of these agreements is rather time consuming.”

Many end-user organisations make the misjudgment that purchasing software as a service (SaaS) through a single vendor will eliminate the need to manage multiple clouds. But, says Justin Parks, global head of online services at Getronics, the reality is more complex.

“When you purchase SaaS, you need to remember that your vendor may well be purchasing its own cloud infrastructure from a third party,” he says.

For your vendor to meet your SLA, it will need to ensure that its own cloud provider is delivering on contractual obligations. Only if this third party is providing a robust cloud infrastructure can your vendor keep services up and running. “If your vendor has no come-back on its own cloud provider it may be rendered unable to meets its contractual obligations to your business,” says Parks.

Parks adds that this type of multiple cloud management is largely hidden - many IT professionals do not realise the need to manage up the cloud supply chain.

“It’s important to consider end-to end service delivery, gaining visibility into SLAs across the supply chain to ensure you can rely on delivery of the services being provided to your organisation.”

Another problem to be aware is who is responsible for what. “Where there are multiple clouds and multiple providers, the boundaries of responsibility (provider to client and between providers) become very blurred,” says John Sheridan, head of ITO practice at independent specialist outsourcing advisory firm Alsbridge.

This, coupled with more complex service management and integration challenges means that the operating model will inevitably change and much greater clarity of ownership and accountability will be needed. “This is further compounded when cloud providers operate under different service levels and pricing regimes making it increasingly difficult to effectively manage the commercial dynamics of an ‘on-demand’ model,” says Sheridan.