Forecasting the Future of Cloud
Most of us are intrigued with the current cloud and data center landscape and how it will change in the near future. We have seen three fundamental changes that I believe will shape how and where infrastructure services will be delivered and consumed:
Late stage adopters emerge - There is a new addressable market for data center and cloud based services: the laggards. The name sounds negative, but many of these organizations had sound reasons to resist moving applications and workloads out of their four walls and into a third-party provider. In-house data centers make sense for organizations who have robust IT staff, very low cost of capital, just-in-time supply chains with a low-cost basis and don’t experience unpredictable swings in compute and storage capacity. For that set, moving applications to a colocation or IaaS offering is not feasible and return on investment analysis supports that notion. That said, the numbers of those organizations are shrinking by the day as business transformation trends such social, mobile, cloud and IoT based functionality require data to live closer to the edge.
The tiers of cloud collapse on each other - Most industry and research organizations like Gartner, IDC and 451Group loosely follow the NIST (National Institute of Standards and Technology) definitions and taxonomy. To paraphrase NIST’s cloud definitions, a cloud service must be self-service, on-demand, support resource pooling and have ubiquitous network access. NIST also refers to three primary deployment models; Infrastructure, Platform and Software. These characteristics and models were ground breaking back in 2009 when NIST first published their taxonomy, but buyer behavior and technology evolution has redefined the cloud services market. For starters, public cloud compute services (virtual machines and associated storage) are being procured in a scheduled or reserved manner, rather than on-demand. Cloud users have seen a 40-50% discount over on-demand instances when they can forecast when they will need extra capacity. Secondly, Enterprise adoption of cloud services has been focused on both Infrastructure and Software as a Service; platform services have failed to gain critical mass. This is not an example of organizations failing to use the cloud for application integration and development, it’s actually quite the contrary. DevOps has become the tail that wags the dog. Developers are one of the largest segments of cloud users, and they just don’t see the value of having a distinct IaaS and PaaS provider. We believe this will force the collapsing of the tiers and Infrastructure as a Service providers will envelop platforms services to deliver more application integration and infrastructure stacks.
Data center leasing not slowing down - Most of us who have been following the IT services markets over the past 20 plus years have seen 4-5 cycles of boom and bust around data center facilities. Boom cycles force both enterprises and service providers to build more facility capacity to support compute and then when markets bust, facilities consolidate and potentially shutter. Enter the hyper growth of the cloud services market. As most research indicates, cloud providers are the largest leasers of data center capacity and show no signs of slowing down. So, whether enterprises host a majority of data in dedicated servers or in the cloud, the data center will ultimately continue to house that infrastructure.
The conclusion that we draw from these trends is that the adoption of cloud services will continue to drive capacity to data centers of all styles, sizes and locations. The specific workloads that can live entirely on a customer premise will start to be retired as the hybridization of IT continues to thrive.