Victoria-based Peter MacCallum Cancer Centre is to replace its eight-year-old high performance compute (HPC) and storage infrastructure.
The public hospital’s HPC is used by 700 staff primarily for research, but also for clinical services such pathology.
However, the majority of the HPC’s Cisco blade and rack servers are past lifecycle support.
Meanwhile, its storage infrastructure, which uses IBM’s Storage Scale, is currently being hit by “multiple short outages”, according to a request for tender.
“While the duration of outages is normally quite short, this causes major disruption, especially to long-running HPC jobs,” the tender stated.
“Current Storage (and to a lesser extent servers) require significant effort to manage and maintain, reducing capacity to focus on other tasks.”
As of December 2024, the storage solution holds 4310TB of primary data, which is expected to increase to approximately 5000TB by July 2025.
According to the tender statement of requirements, approximately 60 percent of current data is genetic data used for bioinformatics analysis.
The HPC uses one on-site primary data centre and an off-site secondary data centre at St Vincent’s Hospital, East Melbourne.
Peter Mac also manages a Microsoft Azure environment, for which it is currently upgrading its Azure region links. This work is expected to be completed before June 2025.