CHGI Transition

To meet building code requirements, the CHGI Data Centre (DC) in HSC B151 must be decommissioned. To facilitate this change, the storage and compute capabilities housed in the DC are being transferred to the University of Calgary's High Performance Computing site managed by Research Computing Services.

Background

The Center for Health Genomics and Informatics (CHGI) is an initiative that provides a wide range of next-generation genome sequencing services and access to high-performance bioinformatics for sequence analysis to all University of Calgary researchers. New researchers are provided a basic level of networking and storage to make use of CHGI services but also have the option of purchasing additional servers and storage to integrate within the CHGI network to upgrade their usage capabilities. Many of the pieces of the CHGI network equipment have been purchased by researchers and institutions.

The CHGI in collaboration with RCS is working on a planned transition of CHGI data, services and workflows to infrastructure managed by RCS. This will allow CHGI researchers to focus on core duties instead of IT services, and open up the potential for scalability in the infrastructure, and standardize a number of common services to IT’s offerings.

Project Objectives

• Develop a transition plan for the equipment and technical services CHGI offers
• Define ongoing maintenance governance, processes, data privacy, retention and other related processes
• Transition data, users and workflows to IT/RCS managed infrastructure

Approach

The project team interviewed 15 Principal Investigators and researchers in the Analysis phase to understand the current use of the equipment and services provided by CHGI. This engagement took four weeks during which we interacted with the CHGI manager, System Administrator and PIs that leverage CHGI’s equipment for their research. An important aspect in the interviews was the investment made in equipment.

The project team took the information available and discussed the technical options based on current IT’s practices. The equipment, services and workflows have been categorized into 2 groups which was determined based on their ease of transition: Quick Wins, and medium to high complexity.

We identified quick wins that can be used as test cases to build trust with the CHGI user base. The Quick Wins are CHGI equipment, workflows and services that should be easy to transition into existing IT/RCS infrastructure, workflows and services with very little impact and costs. This group of equipment are off or almost off warranty, and also have low dependence on storage. These Quick Wins will be the first step in the implementation.

All other clusters and equipment providing web related services or heavily dependent on the IBM Storage (Medium to High complexity transition) will be transitioned using the option described as Option B – PLANNED 12-MONTH TERM TRANSITION to IT.

Option B keeps the equipment and workflows in their current location of HSC B151 providing CHGI researchers a schedule to integrate their processes within IT infrastructure prior to their current warranty running out.

Quick Wins workflows and data are transitioned into ARC following a defined schedule. At the end of 12-months UCIT will start looking at transitioning the medium and high complexity workflows and data.

Two years into the transition, a decision whether to transition the remaining equipment physically to IT needs to be discussed.

To address the inevitable problem of equipment falling off warranty and funding ceasing, clear communication to the equipment owners and researchers will offer two options if they wish to continue operations. Researchers can choose either to transition their workflows onto existing RCS offerings in ARC, or to purchase new equipment under RCS guidelines. Engagement and collaboration with RCS in this stage will assist in purchasing new equipment that can be integrated with the rest of the RCS infrastructure. These services and workflows will then be transitioned into standard RCS infrastructure, and the new RCS administrator will manage this by following the service levels described in a signed Operating Level Agreement (OLA). After the workflows have been transitioned, the off-warranty equipment would be decommissioned.

Once the first wave of transition happens, a revision to the plan will be made to decide whether to forklift part or all of the remaining equipment to an IT data center.