Technology Overview
One Patient, One Record - Accelerating Healthcare Innovation
Technology Overview
How big can it get ?  How small can it be to start with ?  How will it grow with me ?

Performance & Scaling

Predicting the evolutionary growth of any given HIE - let alone the landscape as a whole - is difficult and uncertain. An effective HIE platform must enable evolution and growth. Scalability is typically discussed in terms of how large a workload a given architecture is able to support for a reasonable cost. However, a pilot project must get up and running to have a fighting chance to grow someday, so an HIE platform should also consider the ease and efficiency with which the same platform can be deployed on a small scale and then grown as needed.

An HIE may have a lifecycle beginning with a pilot phase with limited scope, growing to a larger implementation encompassing several participating institutions, and eventually sharing information across multiple HIEs and many institutions. An ideal system would support a small scale deployment, allow for maximum reuse of initial hardware investments as the system grows, and deliver adequate performance to the full scale system.

HIEBus™ components have been designed and built to both scale up and out. These components utilize a tasking and queuing architecture that effectively quantizes the work done by the system. Components operate with a configurable number of threads on task queues to allow for scaling up on a single multiprocessor machine. Additionally, multiple instances of these components can operate on multiple machines, therefore allowing for scaling out. This allows independent tasks to be processed in parallel.

Given this design, HIEBus™ can also "scale down". It can be fully deployed with a minimal hardware footprint. Because components do not have conflicting resource requirements, all components can successfully be deployed together on a single hardware footprint.

Pilot Program

A small pilot program could be implemented by having the HIEBus™ components (the Record Locator and Security Token Service) hosted at one of the participating institutions on the same machine as that institution's adapter. Any additional institutions would only be required to host a single machine running their adapter.

Full Scale Deployment

A HIE with a significant number of participating institutions will need a dedicated set of machines to host both the STS and RLS each of which supports scaling out across multiple platforms to evolve with the performance and growth needs of the project.

One Patient, One Record - Accelerating Healthcare Innovation
privacy | site map | contact us | site credits   All Content ©2006 CareEvolution. All Rights Reserved.