Server migration has long been a nail-biting exercise. Even the simplest server migration can run into hurdles and the consequences for enterprises can be severe, hampering customer experience, reputation and ultimately revenue.
Thanks to the fast pace of technological change and the will of businesses to cut costs and adapt with agility, there’s no getting around the fact that server migration projects are now unavoidable.
But, oh, the horror! Server migration can go terribly wrong and the internet is chock-a-block with stories about migrations gone awry. MySpace, for example, recently admitted to losing 12 years’ worth of music uploads. While this was good news for people who were in cringeworthy emo-bands during school and couldn’t remember their passwords, it was, no doubt, uncomfortable reading for anyone responsible for data.
The benefits of server migration, on the other hand, can be great, often outweighing the risks, and companies that do it right have the luxury of newfound efficiencies, which more than justify the time and capital spent on these projects.
Deciding where to migrate
Before getting into how to mitigate risks, it’s worth clarifying some of the ways server migration has evolved and why it’s important for organisations to understand where they should move.
A fundamental prerequisite to answering this question is to determine from the outset which types of applications and technologies are best suited to specific cloud services and which are more suited to traditional IT environments or can span across hybrid platforms.
Server migration does not mean ditching servers
While cloud technology has offered a huge leap for IT infrastructure and application delivery, on-premise servers still play a crucial role in the day-to-day running of a business. Here are a few common reasons why:
- Application support: Some enterprise applications are simply not yet available or compatible with cloud platforms;
- Remote locations: Some areas of businesses are located in areas where business-grade networking services are unavailable;
- Regulatory compliance: Some businesses operate in countries that do not permit storing data abroad.
At the same time, on-premise mainframes are extremely relevant too, according to figures, 70% of Fortune 500 enterprises still use mainframes for their most crucial business functions. This is because for some processes such as high-speed transactions, finding an alternative cloud solution on par in terms of speed, cost-effectiveness and capabilities around managing high volumes of data can be hard to come by.
Given servers and mainframes are still relevant today, enterprises have to carry out necessary upgrades. According to research by IDC, server performance erodes by 14% annually. By the fifth year, a server has only 40% of the performance it had when it was new. Modern systems can also reduce energy and cooling costs, data centre space requirements, maintenance time and licensing costs.
Optimising on-premise servers: containers vs PaaS
While PaaS solutions are typically thought of as cloud-native, some PaaS offerings can be rolled out on on-premise architectures. Once the software is installed on one or more servers, the PaaS software arranges the applications into a single hosting platform.
Containers, on the other hand, are standard units of software that package up everything an application needs to run, such as configuration files and dependencies, so applications can runs quickly and reliably from one computing environment to another.
How and why containers should be the way forward for enterprises
Containers build on the legacy of virtual machines but allow for a greater density of applications to host ratio.
For advanced technical teams a container service may be better than a PaaS framework for the desired balance between developer productivity, breadth of viable application architectures, IT operations control and the complexity of implementation.
Container services are also a relatively inexpensive acquisition alternative to PaaS frameworks.
Mitigating server migration risks
While it is vital for enterprises to move away from legacy infrastructure, be it migrating to the cloud or a modern server, all projects must be approached with caution.
“All the risks come about in a server migration when there are unknown unknowns,” said Barney Taylor, MD, Ensono. “It’s actually through the process of doing the migration issues suddenly pop up.”
According to Taylor, rather than thinking about migrating the servers, enterprises should be thinking about migrating applications. Server-based applications are not contained in a neat little box; inevitably, you’re going to have applications that run on multiple servers.
“It’s all about understanding dependencies,” he explained.
5 key considerations for a successful migration
This is where discovery tools come in handy. Discovery tools help to build a map of how your applications run and help identify where there are crucial links and where things might need to be physically co-located next to each other. In other words, it’s about determining whether applications will be compatible with the new infrastructure or not.
Likewise, enterprises should consider running a trial migration to ensure ongoing software compatibility.
Server downtime is another important consideration. How enterprises approach it depends on the nature of their business. If it simply can’t tolerate any server downtime, it needs to safeguard operations with disaster recovery and backup initiatives. Enterprises could also set up a temporary private or hybrid cloud to keep critical processes running during the migration.
It is also essential for enterprises to question if they are licensed to move applications. There are vendors out there who have varying levels of restriction on their software, meaning enterprises looking to run their data through their systems might end up having to make far more extensive licensing agreements.
Finding the means to migrate
While it’s common knowledge that there’s a major talent shortage in IT, nowhere else is it more pertinent than in on-premise IT infrastructure projects. According to research from 451 Research, 29% of organisations surveyed find it difficult to find qualified data centre and facilities personnel.
Compounding the problem, millennial IT workers are coming up short on COBOL and other critical mainframe and server expertise. Despite its rich relevance to today’s business operations, university courses have dropped mainframe and its supporting code, such as COBOL, from the curriculum.
How to choose a Managed Service Provider
As such it’s probably wise for enterprises to assume that they do not have a sufficient level of internal talent to cope with server migration.
However, taking into account the complexity of some migrations, it’s not just a simple case of outsourcing. Instead, the right thing to do is work with agile specialists and service providers who can oversee these complex transitions.
Enterprises should look for people to come in and extend their team rather than replace it. Finally, it’s not about getting server migration done cheaper, it’s about getting it done right.