Life After End-of-Life: Software Lifecycle Challenges and Opportunities

Kurt Hildebrand

Set the Way Back Machine for July 13, 2015, the day before Microsoft officially shut off extended support for the wildly popular Windows 2003 Server operating system. On that day, it was estimated that as many as 9 million servers were about to lose vendor support, including security patches and hotfixes. 175 million websites, nearly 20% of the World Wide Web, were soon to be running on an unsupported, and increasingly unsecure, platform. Many IT solution providers remember 2015 and 2016 as years of chaotic, last-minute, costly migrations. Many customers continued to run Windows 2003 well beyond even the extended support deadline, incurring in some cases massive expense in additional support payments.

Lessons Learned?

2020 is set to be another year of massive transformation in the IT industry, and some solution providers see a “perfect storm” brewing with end-of-life deadlines looming for multiple popular operating systems, hypervisors, and business applications. In particular, Microsoft Windows 2008 Server will lose extended support as of January 14, 2020, along with its partner client operating system Windows 7. Additionally, VMware ESXi 6.0 will be ending general support as of March 20, 2020, along with a host of related VMware vSphere Suite component products throughout the course of next year. On the business applications front, Microsoft SQL Server 2008 will reach end of support in July 2019, and Microsoft Exchange 2010 will officially be put to rest in January 2020.

These software lifecycle realities will be particularly challenging for customers that have not been keeping up with evolutionary developments in data center software and infrastructure over the past five years. In some ways, the Windows 2008 generation of platforms represents the last iteration of what we might start to call “legacy” or “traditional” models, meaning that they were originally built on a “pre-virtualization” codebase and in a “pre-cloud” mindset. The bad news? Making the leap from traditional IT to highly-virtualized, software-defined, and cloud-enabled in one step will likely be daunting for many customers. The good news? The benefits of embracing data center transformation far outweigh both the efforts required in making that leap, as well as the costs and risks of inaction.

Navigating the Options

The three most common approaches to dealing with loss of support for a key workload or application are:

  • Upgrade the operating system
  • Migrate to the cloud
  • Re-platform the application

The most traditional approach to dealing with operating system or software end of life (EOL) is to simply upgrade the software to a supported version. Modern operating systems such as Windows Server 2016 are built on modern software-defined features, contain cloud capabilities, and will be supported well into the 2020s. However, legacy applications may need extensive modifications in order to run properly or may not be compatible with new operating system versions.

Another popular solution is to move workloads to the cloud. In some cases, moving an application or the entire operating system to a public cloud environment can put it into a supported configuration, or at least shift the end-of-support problem to a service provider. This solution has a lot of potential benefits but requires extensive research and testing to ensure that applications are appropriate candidates for a cloud computing model, and that costs, performance, and security will be as expected.

The most cutting-edge or forward-looking method to dealing with software end-of-life is the idea of re-platforming or re-factoring the workloads. Re-platforming is the concept of extracting the core components of the application and porting to a new platform—which could be a different operating system, a different database, a container-based platform, or a cloud-native infrastructure. Re-factoring by contrast is an even more advanced process of extracting the underlying functions of the application and re-architecting them for a dramatically different platform construct, typically rearchitecting legacy applications to run natively in the cloud or in containers.

Connection recommends that customers follow a standard methodology for evaluating workloads on a case-by-case basis to determine the most appropriate migration path to meet the business needs for that particular application. A one-size-fits-all approach will likely result in some type of problem or missed opportunity—for example, spending time and resources trying to force an application down a path it was never meant to go, or alternately making a quick decision because “it would work” but missing out on a much more cost effective or efficient solution.

End-of-Life Planning

These are the steps we recommend you take as your software approaches its EOL date.

  • Assess your current state: It’s critical to get an accurate picture of the current state of your environment. Update your documentation, perform a fresh inventory, and map out system dependencies before starting research on potential upgrade paths.
  • Research your options: Don’t necessarily take the first option that comes along. In many cases, there will be more than one way to approach any given workload. Compare various models and strategies and begin to group applications together either by dependency or by platform type.
  • Plan your strategy: Take your research and put specific details in place for each group of workloads. Build a matrix for upgrades, cloud migrations, and re-platforming opportunities. Create a plan to include inter-operability requirements and rollback plans in case any migrations run into issues prior to the deadline.
  • Implement the plan: Ensure that you have clearly defined success criteria for each workload. Perform proof-of-concept tests and validate prior to migrating production workloads.
  • Avoid the next cliff: Implement a software lifecycle management strategy. Work toward a more iterative update and upgrade methodology. Consider the advantages of planned 12-, 24-, and 36-month software upgrade cycles.

No Problems, Only Opportunities

There’s no magic bullet solution for dealing with EOL, whether your personal situation is around Windows Server, VMware ESXi, or Microsoft Exchange. The current state of information technology is more complex than ever before, but there are more opportunities to solve business challenges in new and creative ways. Embrace the change, evaluate your options, and consider working with a partner to help navigate or accelerate your transformation.

Kurt has over 20 years of experience in storage, networking, virtualization, clustering, disaster recovery, and business consulting. He holds several professional certifications, including Citrix Certified Enterprise Administrator (CCEA), Cisco Certified Network Associate (CCNA), Certified Novell Engineer (CNE), EMC Certified Integration Engineer (EMCIE), Microsoft Certified System Engineer (MCSE), VMware Certified Professional (VCP). In his free time, Kurt enjoys strategy games, the outdoors, theater, and live music.