Government Health High Performance Cloud Computing Capability

High Performance Computing

High-performance computing (HPC) is the use of super computers and parallel processing techniques for solving complex computational problems. HPC technology focuses on developing parallel processing algorithms and systems by incorporating both administration and parallel computational techniques.

High-performance computing is typically used for solving advanced problems and performing research activities through computer modeling, simulation and analysis. HPC systems have the ability to deliver sustained performance through the concurrent use of computing resources.  The terms high-performance computing and supercomputing are sometimes used interchangeably.

High Performance Cloud Computing (HPC2)

High-performance cloud computing (HPC2) is a type of cloud computing solution that incorporates standards, procedures and elements from cloud computing. HPC2 defines the techniques for achieving computing operations that match the speed of supercomputing from a cloud computing architecture. In this context, HPC cloud computing will provide data scientist with access to a massive pool of highly reliable and redundant computing infrastructure that can be provisioned on request and released when not required. The complete solution may include storage, hardware and application software, all of which will be delivered through cloud on an on demand basis.

Transitioning to HPC2

As industry, government and non-government organization advance their digital transformation the value of high performance cloud computing is inspirational.  Many companies will say, “yes we do cloud based HPC2”, yet asking what percent by computing, storage and number of projects and FTEs the number becomes about ~10% based on industry benchmarks.  HPC2 is inspirational when you compare before the cloud to in the cloud performance increases significantly.  Measuring the impacts is well documented and most cases measure throughput in number of concurrent jobs or processes and time to see results.  Benchmarks anticipate increased throughout for advanced analytics projects by 3-5 times and improved time to see results are 10-20-30 times faster than prior.  Inspirational is certainly a good description a 5x increased throughput or a 30x faster processing is transformational. 

HPC2 Platform

Data models are many and we approach client infrastructure as transformational as well.  Many are new and want to try or explore as an approach to gain leadership support to have more resources to implement with the ultimate goal to change the infrastructure.  For the most innovative, a broad transition in total with training and development integrated then it becomes a new way of working.  Systems and processes can change, however, the talent may need to be upskilled, trained, repositioned added or reassigned.  All these considerations we have addressed and taken into place.  To support our experience and past performance we describe the situation, challenges or complications and our solutions, results and impacts.  Certainly, every client, customer, project and prior to HPC2 varies but the framework to scale implement and bring the client along have similar approaches, cloud platforms and data models.  With that said we speak to HPC2 as specific to transition to HPC2 but are flexible with the clients prior HPC or non-HPC and transformational HPC2 partners (i.e. Google, AWS, Azure, IBM, etc.).

 Past Performance

HPC Past Performance:  Industry / Commercial / Non-government and Non-Government Organizations

 

Government Agency	Complication	Solution and Results
Government agency (NIDDK) intent to expand awareness, utilization and applicability of clinical data in 3 repositories.	Data within the databases was HPC or cloud based but were not able to be integrated and were a per use export per request.	Solution proposed HPC2 with patient tokenization (Healthverity) to connect the NIDDK to broader databases in the USA.  An open ecosystem of the database

 

Government Agency	Complication	Solution and Results
FDA CTP needed to be more predictive to improve OCE’s efforts to avert violation risk by retailers.	Developed and proposed to FDA an HPC2 statistical algorithmic model evaluating violation risk and other public health factors to inform the retail inspection program enabling the Federal Food and Drug Administration (FDA) Center for Tobacco Product’s (CTP) Office of Compliance and Enforcement (OCE) to obtain data results.	Solution is proposed FDA CTP OCE who is considering solution for consideration. Preliminary impacts was seen that c-stores within a certain geographic area near a high school sell a higher rate of tobacco products.  Helping OCE to support a rank or list of stores to both inspect for violations or adapt policy.

 

Government Agency	Complication	Solution and Results
FDA seeking to address the USA supply chain risk for healthcare, pharmaceutical, biological and medical device products, active pharmaceutical ingredients and components.	Develop and proposed HPC2 USA Pharmaceutical & Hospital Products Supply Chain ecosystem to estimate available supply for life saving medicines across the USA health system due to COVID.  Insights quantified the available supply of hospital products when peak COVID active patients impact inpatient; i.e. when USA reach x# of COVID inpatients how long will the pharmaceutical supply of products last.	Solution is proposed FDA who is considering solution for consideration.   Preliminary impacts was shown that when the USA population reaches a certain threshold of ICU meds certain medications begin to run out of stock with delayed re-stock given lag in the supply chain.

 

Government Agency	Complication	Solution and Results
HCP2 solutions proposed and awaiting feedback at:   NIDDK Repository FDA CTP OCE NICHD Maternal Morbidity DOD DHA Navy Medicine Data Science NIH Public Health NIAID Genetics SBIR Women’s Health CDC FDA COVID HealthVerity DTRA WMD	HPC solutions and non-HPC solutions requiring scale to HPC2	Solution is proposed agency who is considering solution for consideration.