dcn-na logo
Story image

When the cloud falls to earth, it could be time for cloud repatriation

04 Mar 2020

Article by Intel Data Center Management Solutions senior application engineer Rami Radi.

For many of today’s applications and workloads, cloud computing offers the enterprise a host of advantages over traditional data centers, including lowered operational and capital expenditures, improved time to market, and the ability to dynamically adjust provisioning to meet changing needs globally. Consequently, there has been a massive shift to cloud migration over the past decade, with cloud computing trends showing significant year-over-year growth since it was first introduced, and Cisco predicting that by 2021 cloud data centers will process 94% of all workloads.

According to MarketsandMarkets, the global cloud computing market is projected to surge at a compound annual growth rate (CAGR) of 18 percent to reach approximately $623.3 billion by 2023, up from $272 billion in 2018.

Today, however, we are seeing more companies bringing workloads back into their data centers or edge environments after having them run in the cloud for several years because they didn’t originally fully understand their suitability in a cloud environment. 451 Research has referred to this dynamic as “cloud repatriation,” and a recent survey found that 20% of cloud users had already moved at least one or more of their workloads from the public cloud to a private cloud, and another 40 percent planned to do so in the near future.

All of this begs a deceivingly simple question: “How do I know when a workload would be better off running in or outside of the cloud?”

When latency, availability, and control are key

As with any IT decision, an inadequately researched, planned and tested process is likely to cause setbacks for enterprise end-users when the organization at large is faced with uncertainty whether to move an application or workload out of the public cloud and return it to an on-premises data center or edge environment.

Very often, moving an application or workload from the cloud makes good business sense when critical operational benchmarks are not being met. This might mean inconsistent application performance, high network latency due to congestion, or concerns about data security. For example, we know of one Fortune 500 financial services firm that was pursuing an initiative to move its applications and data to the public cloud and only later discovered that its corporate policy prohibited placement of personally identifiable information (PII) and other sensitive data beyond their internal network/firewall. Although many security standards are supported by public cloud providers, because of its internal policy, the financial organization opted to keep its data on-premises. 

Some companies, such as Dropbox, have chosen to migrate from the public cloud to benefit their bottom line. While cost is but one criterion for leaving, it is a major one. In the wake of leaving the cloud, Dropbox was able to save nearly $75 million over two years.

Generally speaking, applications that are latency sensitive or have datasets which are large and require transport between various locations for processing are prime candidates for repatriation. Consider smart cities and IoT-enabled systems, which create enormous amounts of data. While cloud computing provides a strong enabling platform for these next-gen technologies because it provides the necessary scale, storage and processing power, edge computing environments will be needed to overcome limitations in latency and the demand for more local processing. 

Additionally, if your applications and databases require very high availability or redundancy, they may be best suited to private or hybrid clouds. Repatriation also provides improved control over the applications and enables IT to better plan for potential problems.

Yes, moving to the cloud means a decrease in rack space, power usage and IT requirements, which results in lower installation, hardware, and upgrade costs. Moreover, cloud computing does liberate IT staff from ongoing maintenance and support tasks, freeing them to focus on building the business in more innovative ways. And yet, while many businesses are attracted to the gains associated with public or hybrid cloud models, they often do not fully appreciate the strategy necessary to optimize their performance. Fortunately, there are tools to assist IT teams to better understand how their cloud infrastructure is performing.

Demystifying cloud decision-making

No matter the shape of an organization’s cloud — public, private or hybrid —  data center management solutions can provide IT staff with greater visibility and real-time insight into power usage, thermal consumption, server health and utilization. Among the key benefits are better operational control, infrastructure optimization and reduced costs.

Before any organization moves its data to the public cloud, the IT staff needs to understand how its systems perform internally. The unique requirements of its applications, including memory, processing power and operating systems, should determine what it provisions in the cloud. Data center management solutions collect and normalize data to help teams understand their current implementation on-premise, empowering them to make more informed decisions as to what is necessary in a new cloud configuration.

Intel Data Center Manager is a software solution that collects and analyzes the real-time health, power, and thermals of a variety of devices in data centers. Providing the clarity needed to improve data center reliability and efficiency, including identifying underlying hardware issues before they impact uptime, these tools bring invaluable insight to increasingly cloudy enterprise IT environments, demystifying the question of on-premises, public and hybrid cloud decision-making.

Here are some factors to consider when making a decision about embarking on a course of cloud repatriation:

•    Are you wasting money paying for capacity that is not being used?
•    Are you experiencing regular performance and availability issues?
•    Are you required to meet certain regulatory compliance standards?
•    Do your workloads require low latencies?
•    Do you have the IT staff bandwidth to take control of your workloads?

If you answered yes to a majority of the questions above, it might be time to consider cloud repatriation.