By Tim Malfara, VP of Hybrid IT and Cloud Services, Anexinet
The majority of application workloads should reside in the cloud, and most organizations are better off not running their own data centers. That’s not their core competency, and the resources dedicated to maintaining data center operations could be providing more value to the organization by performing other tasks. However, sometimes it makes sense to pull an application out of the public cloud and return it to an on-premises facility.
Why? Some organizations make a move to the cloud in haste due to a perceived need or directive from upper management. Usually this in-haste move is simply a “lift and shift” operation, without taking the time to optimize the workload for running in a cloud environment. The result is that the workload is going to be more expensive and less reliable than when it was running on-premises.
Another reason to move a workload out of the public cloud is data sovereignty. Let’s say you’re offering financial services to your customers, and one of your customers is from a country that requires data to be stored within their boundaries. You might need to move some of your services into a local data center in that country to comply with their regulations.
The final reason is one of dependencies. There are some applications that simply cannot move to the cloud. Often these are applications that run on some type of legacy hardware (such as a mainframe) or need to be close to a physical resource (such as an assembly line). Any workload that is tightly coupled with these applications may perform poorly once placed in the cloud due to latency and throughput issues. Moving the workload back can create lower latency and higher throughput and improve the performance of the application. If you have excess capacity in your data center and you’ve already paid for the equipment, then moving the application back on-premises could also save money.
If you’re looking at the inventory of applications running in the cloud, keep an eye out for apps that are performing poorly or are more expensive than expected. This is especially true of anything utilizing infrastructure-as-a-service (IaaS), rather than the native cloud platform services. It might make sense to rewrite or refactor the application to lower costs and improve performance, but your organization may not have the budget or appetite to do so. Moving the application back on-premises could resolve the cost and performance issues in the short term.
There are several third-party tools that can facilitate the replication of an application from a cloud environment back to on-premises. Generally, they presume that the application is using IaaS, and so the target on-prem will be standard virtual machines. If you’ve refactored your application to use native cloud platform services, such as AWS Lambda or Azure Cosmos DB, then you’re going to have a more difficult time moving the application. It’s best to perform a replication into a development environment first and test the process, then proceed with a production migration. If the application is public-facing, it may make sense to …
The honoree has been a @Telarus partner since 2011. dlvr.it/RNTcY2
January 21 2020 @ 19:35:32 UTC