Cloud and data center technology changes constantly. That isn’t a prediction; it’s a fact. As tech improves and new models and services appear, IT evolves quickly. An important area where the cloud has revolutionized IT is the ability to modernize quickly, reducing costs to customers. More specifically, cloud price-to-performance ratios get better every year, and this year will be no different. Maybe over a year you don’t notice, but over time, the savings build up. From just 2014 to 2017, for example, the cost of running the same set of workloads on AWS has dropped almost 75 percent. It’s a stunning decrease.
This isn’t earth-shattering news to anyone in the industry, though. What’s worth investigating more deeply is how these sharp price declines and performance gains are driving the market and some trends we see developing over the next year. They aren’t just gut feelings; they’re based on real data, detailed analytics and algorithmic analysis. To drive home each prediction, we’ll introduce one statistic and then delve into what it means.
Statistic 1: Recent TSO Logic research showed that more than 80 percent of workloads are overprovisioned.
What it means: Consumption-based cloud models provide a tool against overprovisioning.
It’s a tale as old as data management. An organization buys a server to house its data and applications without really understanding how much compute power it needs to deliver the workload. The result is capacity that sits idle until workloads grow to the point that it’s needed, wasting money and resources.
This problem is expected for on-premises deployments. Organizations are still trying to accurately predict their capacity needs quarters or years in advance and then invest in hardware that can support it. It’s a cost of doing business and leads organizations to investigate the cloud.
Surprisingly, companies bring the same bad habit to the cloud and overprovision there as well. They often try to manually map their on-premises environment to the cloud in the same manner that they determined their on-premises server capacity, leading them to overprovision. A major benefit of the cloud is that you can right-size on the basis of workload patterns and usage, but determining these metrics manually is virtually impossible given the sheer volume and complexity of options. It’s a major benefit of consumption-based cloud models.
Optimizing your compute, storage and services leads to measurable savings—up to 36 percent according to our research—and it’s something organizations just can’t do with legacy on-premises infrastructure. The savings are substantial, so expect consumption-based public- and private-cloud models to see continued strong adoption.
Statistic 2: KPMG predicts that PaaS will be the fastest-growing cloud platform, growing from 32 percent in 2017 to 56 percent adoption in 2020.
What it means: Application optimization will fundamentally change the market.
Clearly, cloud adoption has grown impressively since its inception almost 10 years ago. This growth, however, has largely been limited to where basic infrastructure is housed. The next cloud wave for many enterprises will involve application modernization and optimization, which will drive the next revolution in cloud computing.
Just as organizations have transferred many of their day-to-day infrastructure tasks to the cloud, application maintenance will be the next task that becomes cloud dominated. This shift from infrastructure as a service (IaaS) to platform as a service (PaaS) will have wide-ranging implications.
For example, organizations may move to the cloud but initially continue to run and maintain their own SQL Server databases. Once they’re in the cloud, they’ll have new options available and might look to move to a database-as-a-service platform. All the major cloud providers are offering migration services and large-scale fully managed database platforms that handle all the database care and feeding. These offerings free IT from managing the database infrastructure; the cloud provider deals with it while the customer simply consumes it. The use cases don’t stop there. Everything from facial recognition to machine learning will eventually shift thanks to the PaaS model.
An interesting potential ramification is greater stickiness when it comes to cloud providers. As the application functions moving to the cloud become more complex, customers can build and scale faster; the tradeoff is that moving those functions from one provider to another becomes challenging.
Statistic 3: According to RightScale, central IT teams are looking more broadly at their role in cloud, including selecting public (65 percent) and private (63 percent) clouds, and are deciding or advising on which applications to move to cloud (63 percent).
What it means: More IT efforts can focus on delivering business value.
IT professionals can never rest. We’re constantly break fixing, implementing, learning and reinventing our work. It’s part of what makes a career in IT so invigorating. One task that has been creeping into IT responsibilities is reporting and forecasting cost and capacity, as well as determining what could be offloaded at a lower cost. The reason is that organizations continue to focus more on their main value: what sets them apart from the competition. Increasingly, everything else can be offloaded to service providers.
The care and feeding of servers is a great example. Take any cyberattack—WannaCry, Locky or any of hundreds of others. IT teams that maintain their own servers have spent time and resources dealing with and patching for vulnerabilities at the expense of business-critical IT projects. By switching that infrastructure from on premises to the cloud, such vulnerabilities no longer consume the company’s time, but the time of the cloud provider instead. These providers are geared to solving such problems at scale and often get advanced notice on issues and fixes. They do what they’re best at while your IT team continues to focus on what matters.
Adding in rote day-to-day tasks, along with other activities that cloud providers can more efficiently handle, will continue to free up IT teams to help the business succeed.
Statistic 4: According to RightScale’s 2017 "State of the Cloud" report, cloud users estimate 30 percent of their cloud spending is wasted, and optimizing cloud costs is the top initiative across all cloud users (53 percent).
What it means: Cloud migration becomes a constant undertaking.
Applications, workloads and data don’t just move to the cloud on their own. The process involves developing and proving a business case, choosing what should move to the cloud first and planning the actual migration. This process can be so involved that companies often just analyze their options yearly—or even less frequently.
As cloud decision making becomes a greater priority and becomes more visible financially, analyzing cloud options and moving workloads will become an ongoing process. After all, cloud options don’t change yearly. At this point they change almost daily. Keeping aware of your options and acting constantly will yield cost savings and performance improvements.
If history is any indication, some of these predictions will come true much sooner than any of us expected, while others may never gain market momentum. Even the most extensive research can’t predict the future perfectly; it can help you make informed calls—and I hope these projections help inform some of your decisions. We can predict one thing with absolute certainty: the cloud and data-management market will never, ever be boring.
About the Author
Aaron Rallo is CEO of TSO Logic. Having more than 15 years of experience in major data centers worldwide, Aaron has an insider’s view of the challenges facing IT transformation, which drove him to develop solutions for improving application delivery across environments, including the cloud. In addition, Aaron served as president and CEO of PNI Digital Media, as vice president for Fujifilm, and in senior positions at Microsoft.