Wednesday, September 2, 2009: 3:20 PM
A public health grid will allow users to discover and utilize a wide variety of services to perform needed functions. Each of these services will define what data is required for them to operate, what functions they will perform, and what results can be returned. The normal paradigm for utilizing these grid services will require that a user send data across the grid to the service to be processed and then receive the results back. This process can be performed over secure network connections and utilize grid security services to keep the transfer of data and results as secure as possible. However, even with these security features, there are instances where this paradigm of sending data to the service is not practical. Two use cases have already been discovered; legal requirements that prevent data from leaving state public health networks and insufficient bandwidth to transfer the amount of data needed to utilize a service.
While policies and laws do change, there are states that currently do not allow health-related data to be transferred outside of their state or network except in very specific cases. Public health functions, such as disease surveillance, would not be allowed to transfer patient level data to grid services that resided beyond the state’s firewalls. Additionally, certain services may require access to gigabytes of data every time they are ran. These types of services would require significant infrastructure both at the public health department and the service hosting site. In both these cases, and possibly others, it makes sense to leave the data in its current location and instead bring the service to the data.
Our research will show prototype implementations that bring services to the data in easy to use ways to support public health needs.