Senior Director in Colocation Product Management at CenturyLink, Andy Huxtable, digs down a little deeper into the hybrid cloud and comes up with a list of the best practices and what to look for when choosing a hybrid cloud solution
We’re bombarded with news, features, white papers and tweets about cloud computing. Yes, it’s important but it’s not the whole story. We’re on the cusp of a major shift in IT infrastructure models, which ultimately leads to the dominance of outsourced cloud however, the journey requires some big steps.
The hybrid IT model is at the epicentre of this change in the way IT is delivered and consumed. It enables and accelerates a natural flow of workloads from physical, virtual and cloud services, allowing organisations to choose how, when and where applications and data are managed and stored. Hybrid IT may be comprised of any conceivable permutation of on-premise, managed hosting, colocation and cloud solutions.
The traditional capital and labour-intensive model of IT infrastructure ownership has become unwieldy – it’s days are numbered. Over-provisioning to create headroom has become unsustainable – financially and environmentally – and unjustifiable for meeting short-term or temporary workloads.
Virtualisation can play a key role in improving efficiency and driving down hardware costs. Server utilisation typically ranges from 10 to 30 percent for a “one box, one app” approach and can be boosted to as much as 70 to 90 percent for multiple applications.
However, building a private cloud is an asset-heavy exercise that demands substantial capital investment, skilled resource and ongoing operational expense. So as a growing proportion of servers are now virtualised, moving these workloads to a third party cloud is a logical step to harnessing economies of scale and increasing agility, but as anyone who has done this on an enterprise scale will tell you, it’s not necessarily simple.
Perhaps that’s why, aside from the generic trust issues inherent in outsourcing business-critical infrastructure, the two major inhibitors to public cloud adoption are concern over how much management and control resource will be required, and the anticipated difficulty of integrating cloud with existing infrastructure (Vanson Bourne, 2013 survey of 550 IT decision makers, commissioned by CenturyLink).
What’s more, many organisations are still running production applications on old operating systems that can’t be virtualised. This is often due to dependence on legacy hardware, such as creaking mainframe, mini and RISC-based UNIX systems, the need to host and interface with legacy telecommunications systems, or the requirement to support specialised hardware devices.
The reality is that very few large enterprises were born in the cloud. Most tend to deliver some essential services in-house, while partnering with an assortment of vendors to provide other critical or specialist application functionality. As a result, their teams have to contend with a composite of new and ageing infrastructure spread across multiple locations (some on premise, some hosted by third parties) and wide application sets (some integrated, others not). The drawback of maintaining such a complex estate is that managing the supplier ecosystem alone becomes a significant drain on resource.
But whilst there is no one-size-fits-all technology, the hybrid model is particularly attractive as it allows a natural flow for workloads from physical, virtual and cloud services, allowing organisations to choose how, when and where applications and data are managed and stored.
A hybrid model gives businesses the pay-as-you-grow flexibility to keep pace with changing requirements, without costly step-changes in resource. For example, burstable colocation avoids the need to provision headroom on a ‘just in case’ basis by providing on-demand capacity in the cloud, allowing memory consumption to be scaled up and down from one day to the next. A workload-focused hybrid infrastructure is fundamentally about ensuring the right fit for the specific performance requirements of each application, and ensuring it is optimised within a cohesive framework or hierarchy. This affords IT teams the flexibility to meet their organisation’s individual requirements for technologies, policies and regulations. The hybrid partner provides significant levels of integration – delivering applications and data back to the organisation under a single master service level agreement.
Of course, from an operational perspective, it may be necessary for CIOs to review and adjust the resources, skills and culture within their internal organisation. Their efforts will now need to be focused on service management and innovation, rather than delivery.
For organisations struggling to manage the procurement and outsourcing of IT services, the hybrid model of converged managed infrastructure from one provider is truly compelling. It reduces supplier management cost and complexity while providing the reassurance of single vendor accountability. However, consolidation makes it crucial to evaluate any prospective provider rigorously on several fronts….
Capability - The provider’s portfolio needs to be broad and deep enough to cover the complete spectrum of current and anticipated enterprise requirements – from colocation to public cloud, and alpha-testing to mission critical. A key indicator of commitment and long-haul stability as well as scalability is heavy investment in infrastructure to meet customer demand. This is demonstrated by ongoing evidence of setting up private networks that span the globe, lighting up data centres in multiple locations, as well as continuous improvement initiatives.
Commercial flexibility - Outsourcing should prove cost-neutral or achieve savings through efficiencies and economies of scale. A hybrid partner should offer contract and service flexibility that allows scaling up one service, such as cloud, and scaling down another - say, colocation. What we’re talking about here is adopting a consumption based model. An online customer portal can help internal teams stay in control by allowing them to provision compute resource and storage, monitor workloads and move them around as required – just as if the assets were located in their own organisation.
Independence - If an IT estate is fragmented, vendor neutrality is an important consideration. The hybrid provider’s partner ecosystem should allow plug-and-play as required – even to other clouds, if necessary. Broad telecom carrier diversity should be offered to ensure the right connectivity is available within reasonable timeframes.
Connectivity - Vendors should provide inter-site connectivity – not just on a local or site basis but on a global scale. There should be abundant availability of fibre into the data centre building, and the physical and virtual cross-connect services and access to multiple services provided, either from one building to another or via a private network.
Onboarding and support – a hybrid partner should design precisely the right environment for a specific business – one that balances need for performance and resilience with value for money. The provider should also offer experienced consultants to help plan and execute migration strategies, and an implementation team that ensures everything is running exactly as it should.
Security – The converged service must meet end-to-end requirements for security and privacy as well as performance. One hallmark of a provider’s ability to offer world-class infrastructure and data security is that it is trusted by organisations such as governments, financial and medical institutions, which share stringent requirements to safeguard sensitive data to ensure regulatory compliance.
Today, the majority of organisations still prefer the traditional on-premise, in-house owned IT model. In time, the outsourced cloud will achieve dominance over all other models but until then, IT infrastructures will be a hybrid mix of largely outsourced models and we must embrace that!
About the author
Andy Huxtable is Senior Director - Colocation Product Management at CenturyLink (formely Savvis) EMEA a managed hosting and colocation services with more than 50 data centres in North America, Europe, and Asia, with around 2,500 business and government customers worldwide