Source : enterprisersproject.com
Without planning and oversight, you could not only miss out on cloud savings but also get some surprise bills. Consider these five preventative tips.
The promise of pay-as-you-go cloud computing comes with a catch: If you’re not paying careful attention, you could be in for a nasty surprise when the bill arrives.
It’s not all that different from your utility bills at home: Leave the lights on constantly and your tab from the electric company goes up.
With cloud, the problem of a runaway bill could be exponentially worse, contributed to by many machines and users. Unpleasant billing surprises are especially galling because reduced costs have long been the siren call of cloud computing. This creates a compounding problem: First, organizations assume they’ll save money in the cloud rather than developing a detailed plan for how exactly they’ll achieve those cost savings.
Then that assumption often leads to a lack of visibility and oversight because, hey, it’s automatic, right?
“I think the biggest myth is that going to the cloud will magically result in cost savings,” says Ed Featherston, VP and principal architect at Cloud Technology Partners.
Featherston elaborates on this mindset: “‘Oh, I’m just going to take this workload, I’m going to drop it in [a public cloud], I just log into the console, fire up a couple instances – boom, I’m off to the races,’ assuming the magic will just happen,” he says. “Then they get the bill at the end of the month and go, ‘Oh my God, what happened?’”
We’re here to help you avoid that moment. Even if your hybrid cloud strategy is motivated more by factors other than cost – such as flexibility and speed – you’ll still want to keep tabs on your usage and spending to prevent unexpected surprises when the bill arrives. Here’s a starting point: Don’t believe in magic – or rather, understand that you have to be the magician if you’re going to keep costs in check.
“The magic doesn’t happen by itself,” says Featherson. “You actually need to plan.”
Let’s consider some key building blocks of a strategy to avoid cloud sticker shock.
Put proper governance in place
In many cases, runaway cloud bills happen because there are few (or no) rules in place to prevent them – no guardrails to save a developer from accidentally leaving a slew of machines running non-stop over the weekend, for example, running up your cloud spend in the process.
Putting at least basic governance in place is a smart starting point before any move to a public cloud environment, according to Ned Bellavance, director of cloud solutions at Anexinet.
This includes determining a structure and rules for how teams consume resources, access accounts and subscriptions, and more.
The importance of governance grows as your cloud footprint expands to multiple services and platforms. Cloud sprawl, which almost invariably drives up spending, is a key challenge in multi-cloud environments, but one that can be mitigated with proper governance and accountability.
“Having a cohesive plan around governance before doing a mass cloud adoption can save thousands or more in the long run,” Bellavance says.
Cloud’s great selling points – speed, flexibility, and scalability, to name a few – actually increase the likelihood of an unexpected bill, underscoring the need for effective governance.
“The cloud is deceptively easy to scale and there needs to be governance set up around it before deployments start,” says Scott Morley, principal application architect at OneNeck IT Solutions. “Locking down the environment, so only specific users can add resources, is critical.”
Explore vendor tools to support governance
There’s a difference, of course, between making rules and enforcing them. As a starting point for the latter, make sure you do a thorough review of the tools your vendors offer for creating and enforcing policy, often from within your cloud account dashboard.
“The terminology varies by public cloud, but each one has a set of services that help you govern access, permissions, and spending,” Bellavance says.
He shares an example of a governance structure that begins with the creation of a development account for each development group in your organization. From there, you can tailor rules to the team based on its use cases and needs.
“Within [each] account, you can place restrictions on what services the users in the account can consume, what product SKUs they can deploy, and what type of budget the account has,” Bellavance explains.
There are also a growing number of third-party tools out there for monitoring and optimizing cloud spend.
Morley also notes that many cloud vendors offer some type of billing alert feature, which can ping administrators when you’re approaching a specified usage or spending amount.
“This will provide an early warning, which then allows for review of what is driving the cost as well as adjustments in a timely manner,” Morley says.
Be smart about your data
“Understanding data usage is the number-one thing that will help prevent sticker shock,” Morley says.
By this, he means that you need to make cloud decisions that make sense for your data requirements. Consider that most cloud vendors have an “egress charge” – the cost of taking your data out of their cloud. If you have to pull data back down from that cloud environment to make it consumable to end users, for example, you’ve got a formula for ungainly bills at the end of the month.
Automate as much as possible
“Moving large amounts of data to the cloud, only to have to pull it back down for end-user consumption, is counterproductive,” Morley says. “If data is moved to the cloud, access to the data should be moved to the cloud as well.”
Mor Cohen, CTO at Turbonomic, notes the seesaw some organizations ride between under- and over-provisioning infrastructure. Under-provisioning is a killer in the digital age, she says, leading to outages and other damaging performance issues.
In fact, that scenario is why some organizations move to the cloud in the first place. But this often leads to the problem at hand: Overspending, by overprovisioning.
“Most organizations overprovision,” Cohen says. “In order to avoid performance impacts, they allocate infrastructure for peak demand and add a buffer on top of that.”
Again, this is where one of cloud’s big benefits comes with a downside that needs to be managed.
“The ability to easily provision new resources helps application teams deploy the resources they need when they need them, while also creating an explosion of idle and unused resources that become very costly,” Cohen says.
How do you get off the seesaw? Automate as much as possible.
“In order to avoid this and unlock the promise of hybrid cloud, IT departments need to automate,” Cohen says. “There are simply too many dynamic, fluctuating workloads distributed across complex environments for an IT team to manage manually. There are countless important scaling and placement decisions that need to be made, and mistakes are critical and costly.”
Avoiding sticker shock is far from the only reason to automate: As you increase your use of cloud services, you’re going to hit a wall without automation. “The result of automation is scalability – less effort per person to maintain and grow your IT environment, as Red Hat VP, Global Services John Allessio has noted. “If adding manpower is the only way to grow your business, then scalability is a pipe dream. Automation reduces your manpower requirements and provides the flexibility required for continued IT evolution.”
Cloud spend may need to be someone’s job
Governance and automation, in particular, are powerful weapons against cloud sticker shock, but some organizations will hit a point where the distributed scale of their cloud environments will need dedicated human focus to ensure you’re achieving the goals of your cloud strategy while staying within budgetary requirements.
This is a reason why we’re likely to see the growth of an emerging IT role that blends technical and financial skills. This person’s primary role is to manage and optimize multiple cloud vendors and services.
It’s already on the scene: A recent search on jobs site Indeed turned up open positions with titles like cloud optimization analyst and cloud cost accountant. Arizona State University‘s job post for the latter seeks someone who “will be responsible for monitoring infrastructure costs to ensure adherence to client standards; recommending cost reduction measures to ensure optimal configuration from budgetary perspectives; performing infrastructure billing and optimization activities; maintaining cost information on a daily basis using third-party, or custom tools; preparing forward-looking cost estimations to support client and project needs.”
At a certain point, cloud-minded organizations find that monitoring and optimizing cloud spend is literally a full-time job; expect the trend to continue in larger enterprises, especially with multi-cloud strategy becoming the norm.