A few weeks ago, another cloud service provider, HP, announced that they don’t want to place long-term bets or invest more resources into the public cloud, expressing a renewed focus on private and managed cloud capabilities. We’ve seen similar announcements before (RackSpace, Dell, Windstream, etc.), and this shift in focus isn’t exactly surprising.
Analysts indicate that hosted private cloud and hybrid cloud are seeing the fastest adoption and are in the greatest demand among service providers’ enterprise customers — so much so that enterprises are expected to spend up to 50% of their cloud budget on these types of services. In addition, it’s the managed services on top of infrastructure that will offer the strongest long-term revenue opportunity for service providers. 451 Research advises that every managed infrastructure provider should be preparing to answer these enterprise demands.*
We definitely see that our service provider customers are preparing for this growth in hosted private cloud services. In a recent survey of our service provider customers, we saw that a little more than half initially purchased SolidFire’s AFA for deployment in their hosted private clouds. In public cloud use cases, service providers are driven to us primarily because of our quality of service (QoS) capabilities. They are looking to provide a consistent application experience to all users in a tightly packed multi-tenant platform across mixed workloads.
But in a private cloud, the key feature service providers look for is scale-out capabilities.
This makes a lot of sense. Rapid elasticity (i.e., ability to add or remove IT resources as needed) was cited in an ESG survey of enterprises as the most essential private cloud characteristic. Scale has always been a challenge in private cloud environments, where providers try to balance right-sizing (to avoid waste) with the flexibility and elasticity customers need in today’s digital economy. The applications typically deployed on private clouds are mission-critical, revenue-generating production apps usually with constant, yet variable, data growth. They require the highest service levels, with 100% availability, and can’t easily tolerate maintenance windows that will result in downtime when it comes time to scale. Having a storage system that can accommodate capacity variability and grow capacity incrementally and immediately without customer impact is key.
In a managed hosted private cloud, capacity planning and management is a critical part of the service offered, and customers are contracting (and paying a lot) for it. Having a storage system that scales out quickly and seamlessly allows service providers to remove the guesswork, risk, and headache from capacity-management tasks (not to mention, removes the need for over-provisioning), allowing them to rapidly and easily provision additional capacity when needed while meeting customer expectations and contractual obligations for high-quality capacity management services.
This is not to say that storage QoS is not important for hosted private clouds, because it is (it ranked #2 in our survey). With QoS in a hosted private cloud, service providers can offer dedicated compute with shared storage, delivering the functionality and performance of a dedicated storage array at the price point of shared cloud storage, thereby lowering the total cost of hosting contracts and even getting the ability to offer SLAs for storage performance, which is a great benefit for these mission-critical applications.
Even more important is the mitigation of performance troubles that QoS affords. Again, customer expectations are high in a hosted private cloud. If performance problems arise (and we know they do since performance issues are the #1 reason for support calls into service provider help desks), they need to be remedied immediately to uphold strict service levels around response time and resolution time. QoS can help prevent performance problems from occurring in the first place and gives administrators the ability to immediately address them if they do occur.
In fact, one of our service provider customers removed 100% of performance issues in its private cloud as a result of QoS. With QoS you can support mixed workloads with variable performance requirements on the same storage array, ensuring each gets the exact amount of performance needed. This is precisely the kind of service managed private cloud customers demand and expect from their service provider.
Coming in as the third key storage feature for hosted private cloud is comprehensive data reduction from in-line deduplication, compression, thin provisioning, and space-efficient snapshots, reducing the data footprint by 5-10x. This is mainly an economic factor. Higher efficiency cuts the energy, heating, and cooling costs required by storage infrastructure. The increase in effective capacity ultimately reduces the cost/GB, allowing service providers to offer high performance and availability at a competitive price while achieving greater profitability of their hosted private clouds.
There’s a lot more to say on this topic and how service providers can build and manage private clouds that meet customer requirements while delivering high profitability, so we plan to release another post on this subject.
Stay tuned for part 2, where we’ll talk about the role of automation and orchestration in provisioning, configuration, maintenance, and change management in hosted private clouds.
*451 Research, Managed Infrastructure Market Overview 2015, September 2015