Get Your On-Premises Infrastructure Ready for Cloud Repatriation

The dark lining around cloud services is becoming more pronounced. Recent price hikes as well as challenges in having vendors tailor these applications are forcing enterprises to move their cloud applications back on site. The process can be daunting, so they often need help with the transition.

Corporations adopted cloud to lower costs. In the early days of these services, they realized those benefits. But like other markets, cloud is dynamic, and recent changes are negating its advantages. In cloud’s fledgling years, the market was greenfield (every company and application were potential targets), new opportunities abounded, the market grew quickly, and vendors turned significant profits.

But the landscape has changed dramatically. One pandemic ripple effect was that enterprises invested a lot of money in cloud systems. Companies accelerated the digitization of their internal operations by three to four years. Also, the share of digital or digitally enabled products in their portfolios accelerated by seven years.

The Good Times End

One impact was the cloud boom went bust. The hyperactive market growth slowed. Vendors felt the changes’ growing impact and cut back on staff and services.

Another outcome is cloud pricing is no longer dropping. Instead, it has been stable or rising, sometimes in an insidious fashion. As a result, corporations are taking a close look at what they are spending and find that on premises systems have become much more attractive. In October 2022, Basecamp announced it was leaving the cloud because deploying their own servers was more cost effective.

Cloud Struggles to Serve Complex Workloads

A related issue is a major shift in software design, which also impacted pricing. Applications have become more flexible, which is good for enterprises but not for suppliers. Vendors’ cloud business model works best when applications are very simple. Then, suppliers crank out multiple instances of the same set of services, which eases their administrative work, and pass those benefits onto customers.

The metrics morph when clients add services, functions, and data feeds, steps many have been taking. Growing interest in areas, like artificial intelligence, and Augmented Reality/Virtual Reality, require high performance processing, dozens or even hundreds of GPUs, terabytes or petabytes of data, and special programming tools. So, enterprises now run complex core compute workloads in the cloud.

As the work for suppliers becomes more complex, time consuming, and costly, they pass their expenses on to their customers. To what end? Some companies now spend up to twice as much on cloud services as they did when they first migrated their workloads. Nvidia estimates that moving its large, specialized AI and ML workloads back on premises yielded a 30% savings.

Tough to Loosen Sticky Cloud Data

After deciding to bring its applications back into the data center, many businesses find the process daunting. Moving software and data from the cloud to on-premises is cumbersome. Once applications are in the cloud, they tend to get very sticky. They have many users, different types of data (structured and unstructured), large amounts of information, limited migrating time windows, and varying levels of security. Also, vendors place impediments to the move, such as transit fees, charges for data reads and moving information out of cloud storage; and other deterrents listed in the original contract. In addition, incompatibilities among different servers, storage, applications, and system software arise.

Organizations leaving the cloud also need to take a close look at their Application Programming Interfaces and middle layer software. Software suppliers have been adding more features (and more complexity) to make software more plug and play. But tracking all of the connections can be overwhelming.

In an application, everything needs to be connected seamlessly. When companies start breaking things apart, sometimes the whole system bursts at the seams. Gauging what is happening can also be complicated. Application components synchronize on an as needed basis and therefore at different times: hourly, weekly, monthly, quarterly, and even annually. When testing, enterprises must make sure that they account for all of their connections and timeframes.

Aligning service continuity and Disaster Recovery (DR) plans is another piece in the technology puzzle. With corporations relying so extensively on their systems to perform work, downtime simply is not tolerated. So, the business needs to put strong backup and DR capabilities in place in their data centers

One more step is selecting the right physical infrastructure. The enterprise needs to understand how much processing power, network bandwidth, servers, cabling, and power is needed to support the application. Then, it must select the best equipment options, deploy the infrastructure, and put tools in place to manage it. Another choice is deploying and managing the equipment itself or handing the work over to a third-party hosting specialist.

Finally, businesses face personnel issues as well. When moving to the cloud, IT departments often personnel away from system administration maintenance to areas, like DevOps. When they move back, the personnel requirements morph yet again.

The implications of applications on the physical side of infrastructure are vast and ever-evolving. One cannot overlook the requirements of emerging technologies, especially with the advent of Generative AI. Discover how AI growth impacts enterprise storage.

Outside Expertise Needed

The cloud became a key way of supporting business applications. Recent market shifts negated its benefits, so enterprises are repatriating their systems. But moving information back to the data center from the cloud is a complex process. Whether returning to an owned data center or a leased facility, businesses need help from a capable partner. They need one focused on the development, deployment, and management of physical servers, storage, and networking. For example, Top Gun has experience in making the move to or from on-premises, while most partners are singly focused on moving companies to the cloud and not from it.

Cloud metrics have been shifting. Consequently, companies are looking at moving applications back on site. Increasingly, the process makes business sense but raises technology and management challenges. Corporations need to be sure that they carefully check their applications and map out a viable migration plan. Those steps often mean finding a third party to help with the change.

Moving applications from the cloud back to the data center is a complex process, one where many companies lack experience. Top Gun has experience in making the change and can put it to work for you.

Blog Author Details

Craig Wilson
EVP, Engineering

Top Gun
Craig’s LinkedIn Profile