Princeton University: Domain of One's Own

Large-scale server migration

Higher Education Case Study

Introduction 

This project, led by Jason B. Rappaport, Senior Manager of Cloud Enablement and Collaboration Infrastructure, Enterprise Infrastructure Services, Office of Information Technology, Princeton University and his colleagues, offers a view behind the scenes of a large-scale server migration over a period of six months, from April to November 2024. This case study is based on the successful migration of three cPanel servers that constitute the institution’s Domain of One’s Own project, and illustrates the technology  and collaboration involved.

This case study shows how the collaborative effort between the Office of Information Technology, their campus partners and Reclaim Hosting, ensured the desired outcome for stakeholders across campus, resulting in the successful and timely completion of the project.  

Setting out the scope of the project, Jason explained: 

“This service is being used widely by faculty and departmental staff across three servers, which we refer to as personal, development and production. The migration was essential in order to shore up security and improve accessibility. The migration process also offered the opportunity for us as a team to start the conversation with campus partners and to take stock, and review and streamline about 30% of accounts and sites, both in respect to quantity and quality”. 

The collaborative process identified high-risk sites based on accessibility or outdated technology, which could be anything from outdated plugins, themes and/or PHP versions. This assessment informed the overall risk profiling for sites across the entire project. Typically, site types included labs for faculty or groups of educators undertaking research as well as sites focused on research dissemination. Another group of sites were department sites dedicated to specific academic subjects, showcasing and sharing work, including hosting files and redirects. Also identified were a number of sites set up to host specific applications enabling staff to perform self-service actions. The key here was to foster dialogue between all the stakeholders involved, making time for outreach and communication. 

Such a large-scale, high-impact migration project affecting hundreds of sites shines a spotlight on the work involved in shoring up technical debt to enable thousands of users to work safely on the open web. It provides a unique opportunity to take a deeper look at what it really takes to make technological change management successful and, importantly, the people that make it happen. 


Putting people at the heart of change management

It was clear to the team from the outset that the challenges with this migration process would not lie in the technical domain, but in managing a multitude of relationships and communication via campus partners over a period of months. 

Even before agreeing to the migration schedule, outreach and an information campaign started. This collaborative process at the heart of the change management process took into account the user’s needs, academic context and impact on research. Jason’s colleagues Steve Knafelman and Kyle Jackson were part of the team that led this effort. They explained their carefully considered approach thus:

“Our communication included briefings, outreach via IT and campus partners, and also email communications to groups of users affected. We then tracked responses and carried out individual follow ups where faculty did not respond to group communications.” 

Based on what they learnt through group and individual outreach, a central log of all cPanel accounts was compiled, which listed and checked individually, provided an overview and offered a way to cross-check and track progress for the team. It also acted as a completion checklist to ensure all migrated sites were loading/functioning as expected post-migration. 


What helped the project succeed and shine

A key factor that helped make this large project so successful was consistent communication between the different stakeholders involved. The team ensured clear expectations about what would happen, when and how. 

John Sophy, Enterprise Infrastructure Services, further explained the approach to managing risk during the migration process: 

“During the preparation to the migration the team at Princeton grouped sites into the schedule according to a detailed risk assessment process that took into account PHP version, size and databases, based on which a risk score was calculated and then a balanced group of sites was scheduled for each week. This ensured that a balanced group of sites was migrated each week and avoided too many high risk sites being affected in the same week.” 

Migrating 25 cPanel accounts per week required close collaboration between the team at Princeton and the team at Reclaim Hosting. It ensured smooth and timely communication, which in turn helped manage the risks involved and ensured the delivery of this project on time and on budget.


Key take aways

Reflecting on the success of the project, Jason and his colleagues highlighted what for them stood out as take aways which will help replicate the success of this project for future initiatives:

  • Carrying out robust due diligence proved mission critical and enabled the team to successfully manage the risks involved in migrating such a diverse range of sites including high demand ones; 
  • Focusing on communications, both between the teams and internally within the organization, was essential and individual follow ups helped avoid issues with users of high profile sites.  

Now that the migration project has been successfully completed, work to enhance accessibility and security across the servers can continue as faculty and departments work in their Domain of One’s Own.


The Team at Princeton University

We are grateful to Jason B. Rappaport, Senior Manager of Cloud Enablement and Collaboration Infrastructure, Enterprise Infrastructure Services, Office of Information Technology, Princeton University and his colleagues, Steve Knafelman, Technical Support Specialist, Service Management Office, Office of Information Technology, Kyle Jackson, Technical Support Specialist, Service Management Office, Office of Information Technology and John Sophy, Collaboration Infrastructure Analyst, Enterprise Infrastructure Services, Office of Information Technology, for sharing their insights in this case study. 


Learn more

If this case study has sparked your interest, get in touch and explore our Domain of Ones Own program and Managed Hosting services for yourself. 

Read more