Data In Science Technologies: The Apex Experience in Cluster Management

The broader scale acceptance of scientific and statistical modeling, large-scale media processing, and algorithmic computation is changing the nature, design, and impact of High Performance Computing. Once relegated to large national labs, HPC is now being used from Wall Street to Life Sciences. “The challenge exists in finding the right expertise in this fast growing but complex space. There is an increasing need for senior level expertise to correctly architect and/or manage the system operationally as well as identify and resolve trouble areas as they occur,” says Don Schrenk, Senior Partner, Data In Science Technologies (DST). The Atlanta, GA based firm, DST provides Consulting, Management Services and System Integration in the HPC space to solve these common struggles.

Team DST lays great emphasis over proven methodologies for a wide array of services. From thorough analysis of the clients’ existing workflows to tuning the cluster resources to achieve better performance, DST provides a comprehensive set of services designed specifically for HPC environments. DST is often involved in architecting a new cluster, re-designing an existing cluster or expanding an existing cluster to accommodate new workloads. DST can also assume operational responsibility of the cluster from storage up to the management of the queuing system with their Managed Service offering. Each effort starts with the in depth analysis “During the analysis phase, we strive to nullify the miscommunication between various stakeholders such as the end users, IT teams of the client, and system managers to ensure successful deployment,” emphasizes Andrew Gauzza, CTO, DST. After the discussion with all the stakeholders involved in the deployment and management of HPC solutions, DST outlines a strategic roadmap. “The strategic roadmap encompasses all perspectives of the environment related to application workflow, and infrastructure,” explains Andrew.

As the acceptance of HPC grows DST is able to help organizations reach success much faster. “Often ad-hoc decisions are made at queuing level and utmost precaution should be taken to avoid isolation of the resources to preserve performance,” enlightens Gauzza.

At the end of the day, we want our customer to win

With DST’s Managed Services, the company strives to solve every user need and in the same pursuit has documented the post deployment phase along with the allied and essential procedures to be followed. “We preserve the integrity of the configuration, take care of patches and performance issues, and hold ourselves responsible for providing the services that are needed to keep the cluster operational,” says Don.

The comprehensive methodologies to analyze the riddle in computation, devise the solution, and support the HPC environment post deployment has enabled DST to craft several success stories. In one instance, a client sought the replacement of its existing cluster that was not meeting the performance needs of the scientific community. The collaborative approach and technical design by team DST drastically increased the amount of research that was being performed. DST’s design, management and workload tuning eventually ameliorated the compute time from five hours to one hour fifteen minutes.

DST aims to create success stories through its innovative offerings and customer-centric approach. The company has developed a tool complimenting HPC environments called DataLogger that creates custom and inferred metadata tags and provides a chain-of-custody for data across the enterprise. The development of cutting edge solutions such as DataLogger is the culmination of creativity, customer-centricity, and technical expertise of the leadership team that comprises Willbanks, Schrenk, and Gauzza. “At the end of the day, we want our customer to win,” concludes Schrenk.

Data In Science Technologies

Atlanta, GA

Don Schrenk & Deborah Willbanks, Sr. Partners, Andrew Gauzza, CTO

Provides customized solutions to solve common data management struggles in High Performance Computing

Whitepapers of Data In Science Technologies

Data In Science Technologies