At some point, every enterprise that has moved a significant percentage of their workloads to the cloud comes to the same conclusion: This is harder than we thought. As someone who has been on both sides of the issue, I can attest.
The improvements that flowed so easily in the beginning of the implementation start to bog down. Increased agility and velocity come at the cost of increased management complexity, with the side effect of reduced visibility into costs and cost drivers.
Cloud optimization, as it turns out, is not something you just do once. It’s an ongoing process that needs to be planned out, shaped, reshaped and managed as you go along. If this sounds daunting, rest assured: it can be done! It just takes a bit of time and patience. Stick with it, and you’ll see results. To use a fitness analogy, it’s kind of like slimming down for the summer – and staying in fit condition after the beaches close for the season.
In fact, optimizing your cloud is kind of like crafting an effective fitness program. Not convinced that this analogy works? Here are a few tips for getting your cloud in shape...and keeping it there.
Do an honest assessment
In health terms, this means figuring out where you are, where you want to be, and how much work you need to do to get there. With cloud, it’s all about getting visibility into your cloud operations.
Cloud operations are hard to control, even in the best of situations. As different departments spin up cloud instances at different times, each group struggles to keep track of its own resources. How much capacity are we devoting to each individual cloud initiative? Who has access to regulate each one? Are we scaling resources up and down correctly to meet the demands of the business? How much are we paying for each resource this month? Next month? Or the month after?
To truly get your cloud in shape, you have to know just how flabby it is and what kind of regimen to put it on. Creating dashboards to track resources and evaluate what’s going on department by department, instance by instance, and workload by workload, will give you a baseline to track against, and also get you executing on a solid plan.
Don’t get taken in by fads; think long-term
Fad diets promise quick results on the scale, but the improvements don’t tend to last. The same could be said for band-aid cloud fixes. Turning the dials to shave the costs off one cloud initiative will lower the bill for a month, but there’s no guarantee that the one-time tweak makes sense over a long-term period.
What’s needed is a detailed plan that takes into account both the short- and long-term needs of the organization. Are certain cloud instances tied to critical business goals? Which departments absolutely, positively can’t run short of resources? Are there certain instances that need to be dialed up and down radically to react to spikes in usage or trigger events?
The bottom line is that cloud programs need to be planned out effectively – and then revised according to changing business needs.
Don’t take risks
As any health-conscious individual knows, food carries varying levels of risk. High-calorie, fatty foods expose people to risk not only of short-term weight gain but also long-term problems. Similarly, using improper form or ignoring that old college injury that flares up from time to time can cause serious damage to your body in the long run.
Cloud has its own type of risks. Technical risks can range from open ports to sensitive data that’s unencrypted. Operational risks are usually process-oriented, such as if you have too many permissions or if password policies are being ignored. Both can put the enterprise at serious risk -- and in the headlines for all the wrong reasons.
To limit those risks, organizations need to stay a step ahead of malicious actors and make sure to uncover risks before they become problems. Preventive policies and controls that continuously monitor for vulnerabilities can help organizations know when to take action.
Do what works for you
Judging by the number of weight loss programs pushed by weight-loss “experts,” it’s clear that people interested in dropping a few pounds like to seek out advice. They want a quick-fix solution that gets them in the condition they want to be in – now. Problem is, a weight-loss workout plan should be right for YOU – one that takes your needs and abilities into account.
Same goes for cloud operations. Don’t assume that someone else’s strategies are the best and copy them for your own. Just because a similar organization scaled its cloud operations at a certain pace doesn’t mean that’s the right pace for you. Perhaps you have business needs that require you to keep certain workloads on premises – making hybrid cloud the optimal set-up. Come up with what works best for you and institute your own specific cloud strategy.
Don’t stop when you hit a plateau
Everyone needs a pep talk now and then, in order to maintain their momentum and focus. Fitness initiatives don’t always follow a straight-line progression. Occasionally, workouts fall into lulls, and disciplined diets slide. Cloud improvement projects can follow the same pattern. When projects seem like they’re plateauing, take another look at the strategy and keep pushing.
Find your balance
In an exercise program, the biggest issue is usually lack of follow-through – not keeping to a routine or snacking when you shouldn’t. A fitness program shouldn’t be treated lightly, but sometimes it’s just as big of an issue to overdo it. Too much exercise can lead to injury – and, in cloud, too much provisioning can lead to bloat.
Both under- and over-utilized assets pose a threat. The key is to rightsize your infrastructure to ensure you’re striking the right balance. To truly understand asset performance and utilization, you need to collect, trend and report on both a granular and macro level across CPU, memory, network, disk, IOPS, throughput and other measures. Then you need to evaluate your infrastructure based on application, workload and environment to optimize your performance.
Take control of your health, and be accountable
In fitness, the user takes charge. Only you can pick the intensity of your workout, and only you know how to regulate your own body. But who steps up with cloud? As organizations try to improve their cloud programs, sometimes there can be too many people weighing in. Other times, nobody seems to want to make the call or accept responsibility.
To scale up efficiently in the cloud, you need a “cloud steward” to drive governance. (This was another takeaway from the Cloud Leaders survey.) The cloud steward is a cross-functional individual who is responsible for the ongoing optimization and governance of the cloud infrastructure. This person defines and manages automated policies; analyzes costs, usage, performance and security across environments; and makes recommendations on capacity planning, modeling and forecasting.
Stay in control
Once you’ve gotten your workout plan – or cloud initiative – under control, the trick is keeping it moving in the right direction. In fitness, this can often be solved by just establishing a routine. With cloud, this can be accomplished with centralized governance.
First, you need to align with business policies that define how you want to run your organization while also maintaining flexibility, quick time to market and innovation. Then, automate your processes. If you make the same fix time after time, when utilization dips below a certain point, or costs swell into a certain range, set-up automated triggers to manage the changes.
If you didn’t think the fitness analogy would work, hopefully I’ve convinced you otherwise! At the end of the day, while cloud and fitness may seem like diametrically different concepts, they are similar in many ways. In both cases, you need to figure out what will work best, set a plan and continue to tone to stay in the best shape possible.
Joe Kinsella, CTO & Founder of CloudHealth Technologies
Image Credit: Everything Possible / Shutterstock