A big part of the CIO role today entails hashing out a cloud computing strategy. Deciding whether to build a private cloud or to offload some workloads to a public cloud system while keeping others on-premises is just a start. Then there is the often painful haggling over service-level agreements (SLAs), those crucial contracts regarding the level of service you can expect to receive for your money. But the true test of cloud success begins when it comes time to manage your cloud environment.
The reality is there are far more than a few permutations at play, with enterprises growing particularly omnivorous in their consumption of cloud services, opting to mix and match services in hybrid or multicloud environments. Over 85 percent of enterprises will commit to multicloud architectures encompassing a mix of public cloud services, private clouds, community clouds, and hosted clouds, according to IDC. Moreover, the research firm says that more than 50 percent of enterprises will subscribe to more than five different public cloud services and will continually add, expand, contract, and drop subscriptions based on business needs by 2018. The resulting complexity will make it all the more important to get cloud management right from the start.
The pay-as-you-go nature of the public cloud offers the promise to set it and forget it. But if you’re not continually managing how your company is leveraging the public cloud, you’re liable to get bit.
“A big mistake that many companies make is that they treat, particularly public cloud service, as though it is cable service, where you use it every month and pay a bill at the end of the month,” says Dennis Smith, a Gartner analyst who tracks the cloud management space. “Many find they’re spending more money than they did before [using their on-premises service]. Public cloud providers aren’t going to tell you there are more efficient ways of using their services. You need to manage it similar to the way you’d manage on-premises infrastructure.”
CIOs need to learn to manage those cloud systems with regard to cost, capacity planning, security and other conditions. That need has spawned a modest but growing market for cloud management tools, which companies use to apply policy to as well as automate and orchestrate across public and private cloud services in a uniform way, according to Smith.
Cloud management tools span more than a dozen functional areas, with dozens of vendors in the market offering various applications. But no single vendor offers a silver bullet software system to manage every need. Smith says that by 2022, 80 percent of enterprises will require four to six cloud management tools to rein in their hybrid and multicloud strategies.
To wrangle this growing complexity, CIOs should focus on the following five functional areas.
Cost transparency and optimization: Smith says managing cloud costs is a nightmare for many CIOs, particularly those who are new to cloud adoption. CIOs are used to a world where they purchased equipment, set it up and used it as needed. But they can’t do that with cloud services. Consider this all-too-common nightmare scenario: Developers testing an application sometimes forget to turn off virtual machines Friday afternoon, leaving them running over the weekend even though no one is there to work with the computing jobs. This can cause CIOs to rapidly burn through their monthly allocation for a particular cloud service. Such horror stories are a big reason CIOs are setting up calls with Smith.
CIOs need to closely track and align their consumption of cloud services with how much they’re budgeting for that particular cycle. High interest in cloud cost planning is a big reason why Microsoft last month agreed to acquire Cloudyn, which helps enterprises analyze consumption, enable accountability and forecast future cloud spending. Microsoft’s thinking is that Cloudyn’s software will help make adoption of Azure more appealing to CIOs.
Capacity and resource planning: To ensure efficient operational use of your cloud infrastructure footprint, you need to pay close attention to capacity and resource planning, which is often tied to orchestration and automation functionality. Before switching that cloud service on you need to ask: How much do you need and when do you need to run it?
Security and identity: There are various security-related functionalities needed in a cloud computing environment, including risk management, segregation of duties, single sign-on and key management. CIOs and CISOs need to huddle with their vendors on appropriate cyber SLAs.
Governance and policy: Just as with on-premises technologies, it is essential to create policies that dictate who can access what assets in the cloud, how much and when. Which workgroup or department can access which applications, and from which services can they access them?
Orchestration, automation, provisioning: Orchestration, automation and provisioning capabilities are vital for managing a complex cloud environment. Here, an important facet is “service blueprinting.” Similar to the concept of customer journey mapping, service blueprinting entails spelling out the connections and interdependencies involved in both cloud and on-premises systems.
Of course, the above five tasks may be easier said than done. Smith says that most companies lack the staff to “pull together the tapestry” required to operate a cloud service. “What Lego pieces do I need for that and how do I blueprint it and manage this both on-premises and leverage public cloud APIs and knit together the components that culminate as part of a service,” Smith says. “There’s so much stuff out there that it becomes very, very difficult for one to manage it.”
But it all comes back to cost. In an era of mass digital transformation, CIOs are being asked to reduce run-the-business costs while fostering more innovation. That’s pushed many head-long into cloud systems without forethought as to whether the applications will actually cost less to run in the cloud than on-premises. Sometimes, the cloud, particularly a public cloud service in which the meter is left carelessly running on virtual machines, costs more.
“Not all workloads go [to the cloud],” Smith says. “Those that say ‘in two years, all of my workloads will be in a public cloud environment’ — I don’t think that they have thought through every detail. You may have a large bank that has stringent service levels or serious latency considerations and you may not find a public cloud provider will adhere to those service levels so it’s important to conduct a workload-by-workload, house-by-house analysis to decide what makes most sense.”