Why Cloud Computing and Colocation Go Hand in Hand

Written by Joe Kozlowicz on Tuesday, August 2nd 2016

data center technician installing a switch for cloud computing in a data centerWhile enterprise computing is still hosted in on-premise data centers, the move to hosting providers is accelerating. According to IDC, by 2018, 65% of all IT infrastructure will be located in cloud or colocation data centers. Meanwhile, cloud has the spotlight. Think pieces abound describing the inevitable move to cloud-based infrastructure, especially as on-premise data centers age or undergo consolidation.

Ultimately, cloud computing may mature enough to replace the majority of IT infrastructure loads, but for many years we’ll actually see colocation and cloud working together in concert. Colocation certainly isn’t dead and has its place in the infrastructure strategy of many companies, from smaller cloud service providers to huge enterprises facing compliance or specific hardware requirements.

Is cloud computing the future? It’s certainly a major part of it. But colocation isn’t going anywhere for the next decade or two. If anything, it will continue to grow alongside expanding cloud services.


Colocation Enables Some of Cloud’s Benefits

Colocating servers and other IT equipment offers many of the same benefits as going cloud. While there is still some capital expense involved in purchasing servers, networking, storage, and related items to fill out your racks, you don’t have to spend on the backbone infrastructure like connectivity, redundant power, generators, or the server room itself. That’s potentially millions of dollars in CapEx savings already.

This shift to an operational expense model is a commonly cited advantage of going cloud, but colocation also has some of those flexible spending benefits. Usage-based pricing is also in place, usually billing based on power and/or network use. With a data center provider focused on energy-efficient operations, enterprises can often save significantly on power consumption just by moving current equipment into a data center operator’s facility.


Colocation Will Be Malleable (Like the Cloud)

100% uptime goes hand in-hand with avoiding the CapEx of building a fault-tolerant data center facility. Cloud services and colocation are both expected to just work, with 100% Service Level Agreements often guaranteeing 24/7 uptime.

Because different workloads require different levels of resilience, however, we’re seeing colocation providers offer customizable power densities, redundancy, fault tolerance, and load balancing options. These can be customized based on the facility site as a whole, the data center floor, or in some cases even at the pod or rack level, with a single data center floor offering several different redundancies or power densities as needed.

For example, a website or less-used application might not need 2N+1 redundancy. The company could pay less comparatively to colocate at a service provider site that offers N+1 or even a barebones N setup with no backup power, focusing instead on finding carrier-diverse networks and leveraging load balancing with other data center sites or a cloud-based backup in order to keep services online. Meanwhile, a hospital or an airport might need that 2N+2 resiliency, as a disaster recovery failover might even be too disruptive.


Cloud is Changing the Colocation Game

Colocation providers must work with a more diverse set of IT staff to design the correct environment for future uses. As IT services move more towards an on-demand, as-a-Service model even within an enterprise IT team, the data center engineer is no longer the lynchpin for server rack planning. Cloud architects, IT directors, project managers, and cloud service managers are all involved in data center planning.

Companies mostly aren’t moving wholesale to the cloud, either, and their migration plans usually involve multiple geographically diverse clouds. Colocation therefore becomes a central point to interconnect and offer centralized systems.

Meanwhile, colocation companies are also offering managed services, cloud hosting, private servers, and other added value options. These are often strategic or business driven and must accommodate specific applications or initiatives rather than simply offering the space and power necessary for servers.

Large-scale cloud providers are mostly building or operating their own data centers, but smaller providers need to host somewhere — and in fact, many larger providers also purchase significant colocation space. Similarly, large enterprises have significant existing IT resources. As they begin to migrate specific applications to the cloud, it will be more and more attractive to consolidate or eliminate some data center sites. Eventually, as workloads are split more evenly between cloud and colocation, the currently running on-premise workloads might be better suited to a colocation site than a half-empty data center.


What are your data center plans for the next 5-10 years? Will you keep anything on-premise? Will you keep any physical hardware or kick off a cloud-first initiative? Whatever the case, Green House Data is ready to help you achieve your IT goals.

Chat Now