Solving the puzzle of hybrid cloud [Q&A]
Many enterprises are moving towards hybrid cloud environments, but they face a challenge when it comes to working out how to control their cloud use effectively.
If they fail to do this and govern their cloud use properly, then any gains in agility they achieve will come with high costs and operational risks. We spoke to Andrew Hillier, CTO of Cirba, the company behind the Densify.com SaaS hybrid capacity analytics software, to find out how enterprises can bridge the gap between cloud hype and reality.
BN: What are the main factors driving the move towards cloud use?
AH: There are numerous factors that are driving the move towards cloud use for all sizes of enterprises. Agility is cited the most, as public cloud allows businesses to scale their operations up or down depending on needs, while quickly responding to the business. Cost competitiveness is another. We have even heard from some customers within which boards are pushing to get out of the business of owning and operating data centers. And even if that isn't the case, most organizations want to avoid expanding their physical footprint or building more data centers -- we have one customer that was running out of power in their data center and had the choice of looking to relocate or use the cloud. The reasons are as varied as the organizations using cloud resources.
BN: What's the impact to the business of IT infrastructure in the cloud now being an operational rather than a capital expense?
AH: One of the benefits of old school data centers is that organizations knew how to plan out their costs. They typically over-provisioned to plan for growth and sunk a bunch of money into hardware and software. The move to cloud presents a more dynamic cost structure with variable usage, costs and even changing vendor catalogues and cost structures. That traditional planning exercise is still important, but rather than happening cyclically, it is now an ongoing, dynamic process. This means that everything needs to happen faster, and it requires consistency in approach, with the proper governance and policies in place to not make a mess or have costs spiral out of control. To constantly optimize public cloud OPEX, and to strike the right balance between incurring new costs versus leveraging the sunk CAPEX costs of existing infrastructure, requires more than spreadsheets and smart people. It needs analytics that can incorporate a dynamic model of demand and a deep, up-to-date understanding of the various hosting environments, to scientifically determine the best options for workload placements and resource allocations. We are finding that most companies don't have their heads around how to do this yet.
BN: How important is proper governance of cloud strategy?
AH: All of the benefits of migrating towards a cloud strategy -- flexibility, scalability, security, etc. -- are not worth anything if your IT department hasn't figured out how to properly manage and govern its cloud usage effectively. Proper governance is the most important aspect of a successful cloud migration strategy, but it is often an afterthought in many organizations, as that early cloud adopters in Dev/Test didn't always need to deal with governance rules, and they didn’t experience the long-term costs that can rack up when things move into production. In fact, I've heard surprised reactions from many IT decision makers who began the migration process and were shocked when expected savings turned into higher monthly costs.
To avoid this, companies need to have a formal set of criteria and guidelines for decision making as to where to place workloads in order to effectively leverage both on premise and cloud infrastructure. And this isn't just a documentation and awareness exercise -- these criteria need to be codified into the analytics and automation systems being used, so the right decisions can be made rapidly and automatically. This will help not just in governance and compliance, but also in ongoing cost control. All in all, it's important that CTOs and IT teams remember that the reality of the cloud won’t ever live up to the hype if it’s not being governed and controlled properly.
BN: Does too much control risk loss of agility?
AH: It doesn't have to -- the ability to codify the ground rules and use analytics to optimize and automate cloud hosting decisions means that complex hybrid decisions can be made in real time, giving the business exactly what they need without slowing them down. But don't expect cloud provisioning and orchestration tools to have this level of intelligence -- being good at provisioning a new system is completely different from being very good at figuring out exactly where that system should be provisioned, especially in hybrid cloud. To automate these processes and not create an ungoverned mess, you need analytics that can model all necessary criteria for the workloads, the capabilities of the infrastructure and the governing policies.
BN: What's the future for hybrid cloud?
AH: One thing I think we can all agree on is that cloud is becoming more widely accepted and will play an increasing role in infrastructure strategies. How much ends up in the cloud is anyone’s guess at this point, and will likely vary significantly for each organization depending on the nature of their applications and the patterns of their workloads, but we do know that customers see public cloud as the yard stick for speed and cost against which internal IT departments will be measured. I believe this will be the year that many companies will finally figure out how to leverage that yard stick to scientifically determine what to migrate to the cloud successfully, and what not to. By properly analyzing workloads, many organizations will undoubtedly find that a portion of their applications are more cost-effective and efficient in on-prem virtual environments, at least until their existing data centers run out of space.
And we shouldn't underestimate the power of virtualization -- the rise of interest in bare metal offerings, where you can bring your own hypervisor, is also changing the dynamic. The ability to intelligently stack workloads and over commit resources is clearly very powerful, and also translates to the cloud as a powerful way to drive cost savings. VMware's partnership with AWS is going to enable this, and it could have a major impact on the way organizations adopt 'cloud,' as well as the systems they need to manage and optimize these hybrid environments.