IBM Spectrum Scale Best Practices for Genomics Medicine Workloads
Ulf Troppens, Joanna Wong, Kevin Gildea, Kumaran Rajaram, Luis Bolinches, Monica Lemay, Piyush Chaudhary, Sandeep R. Patil, IBM Redbooks
IBM Redbooks, Dec 22, 2017 - Computers - 76 pages
Advancing the science of medicine by targeting a disease more precisely with treatment specific to each patient relies on access to that patient's genomics information and the ability to process massive amounts of genomics data quickly. Although genomics data is becoming a critical source for precision medicine, it is expected to create an expanding data ecosystem. Thus hospitals, genome centers, medical research centers, and other clinical institutes need to explore new methods of storing, accessing, securing, managing, sharing, and analyzing significant amounts of data. Healthcare and life sciences organizations that are running data-intensive genomics workloads on an IT infrastructure that lacks scalability, flexibility, performance, management, and cognitive capabilities also need to modernize and transform their infrastructure to support current and future requirements.
IBM® offers an integrated solution for genomics that is based on composable infrastructure.
This IBM RedpaperTM publication describes a composable solution with detailed architecture definitions for storage, compute, and networking services for genomics next generation sequencing that enable solution architects to benefit from tried-and-tested deployments, to quickly plan and design an end-to-end infrastructure deployment. The preferred practices and fully tested recommendations described in this paper are derived from running GATK Best Practices work flow from the Broad Institute. The scenarios provide all that is required, including ready-to-use configuration and tuning templates for the different building blocks (compute, network, and storage), that can enable simpler deployment and that can enlarge the level of assurance over the performance for genomics workloads. The solution is designed to be elastic in nature, and the disaggregation of the building blocks allows IT administrators to easily and optimally configure the solution with maximum flexibility.
The intended audience for this paper is technical decision makers, IT architects, deployment engineers, and administrators who are working in the healthcare domain and who are working on genomics-based workloads.