Is it Time for Cloud Repatriation, Cloud Optimization or Colocation?
Is it Time for Cloud Repatriation, Cloud Optimization, or Colocation?
For years IT experts and industry pundits have extolled the cost benefits of cloud migration. The savings potential is real, as many companies who have made the move will attest. But that doesn’t mean it’s the right thing for all organizations ─ or the most cost-effective option, as some of them have learned.
How the Cloud Gets Expensive
When you migrate to the cloud, the cost benefits usually appear early on ─ usually within the first one to three years. (This doesn’t take into account migration costs, which is a separate discussion.)
After all, you’re no longer investing in the resources to run the workloads you migrate. That includes the physical assets, like servers, and the associated costs for maintenance, power, staffing, etc. Because the cloud services provider (CSP) handles the management, maintenance, security, compliance audits, and other responsibilities for the underlying cloud infrastructure, your internal resources are freed up for other purposes – like revenue-building initiatives and innovation.
You’re also trading CapEx for OpEx. And if you’ve been underutilizing on-site servers, you’ll definitely see savings because you only pay for what you use with a cloud model. The cloud’s on-demand-pay-for-what-you-use model also works well when you need to start scaling up to keep pace with growth. The cloud also offers access to nearly infinite storage and computing capabilities for processing enormous datasets, developing machine learning models, and other high-resource usage purposes.
Soon, however, you can find your cloud costs increasing. Those increases can reflect healthy growth, such as growth in your user or customer base, increased digital adoption, or the development of new digital capabilities. But they can also reflect not-so-healthy growth stemming from poor stewardship, lack of resource usage oversight, and cloud waste ─ unnecessary spending on cloud services.
At some point, the cloud can start putting pressure on margins, particularly when growth slows. Some companies have reported that their cloud spending is, on average, 50% of their total cost of revenue (COR).
The economy can also affect cloud costs. With inflation high, last year, Google announced a significant increase in cloud storage costs of between 25 to 50 percent. Data transfer costs for object storage (egress charges) went from free to two cents per GB, even if all the data stays in the same continent. Two cents per GB can quickly add up for something like a typical high-volume microservices architecture, impacting it by hundreds or even thousands of dollars per month.
There’s also the matter of how you handle a cloud migration. Lift-and-shift migrations of existing on-premise applications to the cloud can increase costs if the apps aren’t optimized or remediated to leverage the cloud’s attributes. That means missing out on things such as autoscaling and automated performance management.
The Colocation Option
There are many options available to control cloud costs, from using the cost optimization and monitoring services many CSPs offer to opt for a different IT infrastructure mix – one that takes advantage of the cloud and its alternatives.
Colocation should be among the considerations. Like the cloud, it offers cost savings that come from shared facilities, reduced or no infrastructure investment, and more. You’ll find a lot of good information about why colocation should be part of any IT infrastructure strategy discussed here.
Briefly, colocation offers immediate access to a robust infrastructure. Colocation facilities typically employ the highest standards of network security, including the latest firewalls and IDS systems, to detect and prevent unauthorized access to systems. They usually have redundant network connections and redundant uninterruptible power supplies (UPS) to make sure that business-critical applications always run uninterrupted.
The increased power densities of modern servers also need powerful cooling, which can be very costly on a small scale. Even for those companies with just a few servers tucked away in an on-site server room, the savings from colocation can be significant compared to owning and operating an on-site data center.
Like the cloud, colocation facilities offer the flexibility to jump to higher bandwidth levels to accommodate their traffic demand without making repeat capital investments. Data spikes can be distributed over time across numerous users, helping to reduce bandwidth costs.
The result is IT infrastructure that can expand to support growth with less investment than a private data center. In addition, some studies have even shown that migrating to a private cloud hosted in a colocation facility can outcompete public cloud players on price ─ particularly for data-intensive applications.
Interestingly, a number of companies (large and small) have even started repatriating some of their workloads, moving out of the cloud ─ often to colocation facilities. Among them: 37Signals, the company behind BaseCamp, and HEY. They aren’t alone.
In the 451 Alliance Datacenters 2021 survey, 48% of over 600 respondents indicated that they had moved a workload away from the public cloud to another venue in the preceding 12 months. That’s not to say the cloud was bad, expensive, or anything.
Shadow IT was the top factor for moving workloads/applications away from the cloud (25% of data center/colocation respondents are planning such a shift). Information security concerns were second (23% of those surveyed). Application lifecycle considerations (i.e., different IT environments for test/dev and production) came in third (22%). Regulatory/governance requirements were next (18%), followed by data locality or sovereignty (16%).
Cloud Cost Optimization
Of course, cloud repatriation comes with comes too. But there’s still another good option available for cloud users who want to control costs: cloud cost optimization. There are a variety of cost optimization tools on the market that can help make running workloads in the cloud cost-effective. AWS is well known for the plethora of resources it offers. Most reputable CSPs also offer similar tools and resources for monitoring, tracking, and optimizing cloud usage. You’ll find more information here.
Spotify even implemented a Cost Insights plugin to ‘bring spending data into an engineer’s everyday development workflow.’ The idea was that if they saw data cost at a granular enough level, they would seek out cost optimizations as they would any other optimization.
Your Next Steps
The first step is to determine what you’re currently paying for your IT infrastructure – whether it’s onsite, in a colocation facility, in the cloud (or multiple clouds), or you’re using a mixture of infrastructure. Dig deeper into your expenses. Are you entirely using the resources available? Identify opportunities for increasing efficiencies, optimizing resource usage, and reducing unnecessary costs.
Next, reach out for information. Talk to your current service providers. Ask them for recommendations for optimizing resource usage and cutting costs. If you’re working with a CSP, that provider should be able to help you assess your current IT mix and develop solutions better suited to your needs.
Look at alternatives. Just because you’ve been with a particular provider for a long time (and have become comfortable) or don’t want the hassles involved in change, go ahead and get recommendations or suggestions from other service providers or consultants. Be sure to make a needs/wants list first so they know what your priorities are, what your pain points are, and what kind of solutions will best meet your current and future needs. Cost is likely not the only reason you’re contemplating a change, so make sure to communicate what you want and need.