Working in the cloud will help us run smoother, innovate faster, collaborate more broadly, and reduce costs. But the truth is, most enterprises are spending far more on cloud computing than they had expected. Many IT leaders are wondering why — after moving so many workloads to the cloud — the savings just aren’t there yet.
For the most part, the work being done in the cloud is more efficient than what was happening in the on-prem world. However, many organizations are paying for computing power and storage capacity that they don’t use — or have just simply forgotten about. Capitalizing on the full promise of the cloud requires a planned approach to cloud spending. Here are nine ideas for getting costs under control and building a right-sized and highly utilized cloud infrastructure for your organization:
1. Resolving unused or unattached resources
There’s a reasonably good chance that you are paying monthly or quarterly fees to cloud service providers (CSPs) for capacity that you’re simply not using. It’s a common issue - one of your teams forgets to shut down storage attached to an instance that has been terminated. The CSP doesn’t care that it is not being used; you still get the bill. It’s worth the time and effort to survey your infrastructure, identify unused or unattached resources, contact the CSPs, and have them shut down.
2. Consolidating idle resources
Idle computing also likely consumes at least a piece of your IT budget. A review of your infrastructure might turn up several computing instances with utilization levels that are less than 10%. Paying for the remaining 90%+ in those instances is clearly wasted spending. The solution is fairly simple - identify those low-utilization instances and consolidate them into a single instance. In some cases, people in your enterprise may have intentionally set up these low-utilization configurations, thinking they will need to scale up as workloads increase. But on-demand CSP features, such as autoscaling and load balancing, make this an extraneous expense.
3. Adjusting to peaks and valleys
There are several ways to do this, but readily available heat maps are good tools. There are different versions of heat maps, but they generally give you real-time visibility into enterprise demand for computing capacity. You see when demand spikes and when utilization slows. Heat maps are important mechanisms for cloud cost optimization. Over time, you can typically detect patterns, then align your CSP contracts with the ebb and flow of your usage.
4. Right-sizing your computing services
This sounds simple, but it’s not. When most IT pros talk about right-sizing, they’re referring to the process of matching instance types and sizes to specific workload performance and capacity requirements — at the lowest possible cost. To get to a truly right-sized computing infrastructure, you need to understand the myriad of CSP offerings. This includes evaluating and choosing the right servers, which may be optimized for memory, database, storage capacity, throughput, and other key considerations. Most CSPs want you to get value from your relationships with them, and they have consultants available to navigate the (literally) millions of potential combinations and make good decisions — at least until you get a feel for your options.
5. Prepaying for capacity
Another strategy that requires some math is called “reserved instances” or RIs. If you can accurately predict your minimum usage for a quarter or year, you can purchase (or reserve) capacity at a significant discount. While there’s potential for wasted capacity, the savings generally outweigh the risks. You’ll likely end up adding capacity towards the end of each time-period. The key, of course, is analyzing your historical usage and accurately forecasting changes in demand. Collecting these insights requires a change in behavior for most organizations, as they generally don’t track these metrics — at least not accurately.
6. Leverage multi-cloud
Of course, you can lose volume discounts if you take the multi-cloud path - so it’s not a clear-cut decision. If you do go multi-cloud, there’s also the added overhead of managing multiple cloud instances. But if you measure total cost of ownership (TCO) and factor in concepts like agility in the innovation process and speed to market, and contribution to your innovation process, a multi-cloud approach can be a smart financial decision. In addition to preventing vendor lock-in (i.e., future price increases), you can use a multi-cloud approach to ensure your teams always have access to the capacity and features they need to move quickly and create value.
7. Make strategic tier decisions
Cloud tiering is emerging as an approach to ensuring cost efficiency as you decide how to store your entire range of data, including unstructured data and assets that have minimal use. Local tiers are great for general user files, such as documents and email attachments that need to be retained for 30 days. Remote tiers are generally best suited for 90-day storage. Hot and cold tiers are ideal for general backend data that needs to be stored for up to 36 months. Archive or deep storage tiers are available for data that needs more permanent storage, especially for documents relating to compliance. The tiers that offer the greatest speed and more ready accessibility generally cost more than the deeper storage tiers so that you can save money through careful tier planning.
8. Be precise in your licensing spend
This can be a time-consuming task if you’re trying to do it manually. It’s usually so tedious for some companies that they don’t track their licenses and pay for a great deal of unused capacity. CSPs and other independent software firms offer analytics tools that automate this process and give you visibility into your licenses and usage. If you’ve built a multi-cloud environment with any level of complexity, these analytics tools may be a good investment.
9. Build an understanding of cloud costs
When individual workgroups or departments spin up a cloud instance, they might think the associated costs and potential non-usage are insignificant. Multiplied across an enterprise, however, the costs and the wasted spending tend to be quite significant. Build a culture of awareness: give the leaders of each group visibility into the enterprise’s total cloud spend and regularly communicate the impact of under-utilized cloud instances. This will likely have a positive effect on people who might currently be making costly decisions.
The way ahead
Ask any C-level executive for the top three reasons for moving to the cloud, and one of the answers will be cost reduction. But most organizations haven’t figured out how to control cloud costs yet — especially in a multi-cloud environment. The trade-off at this point is probably good. It’s better to make progress moving to the cloud, then find ways to optimize your spend. Gaining visibility, control and efficiency are going to become important quickly. Out-of-control cloud costs can wreck what should be healthy margins when you bring new products to market. That’s why analytics-based tools and experienced consultants — like Mindtree — can make the task of building cost-effective cloud infrastructures more feasible.
