Microsoft Azure Takes the Search for Kidney Donor Matches to the Cloud
Kidney donor matching has gotten a speed boost with the help of cloud computing. The National Kidney Registry announced that by transferring its matching database to Microsoft's Windows Azure cloud platform, its matching capacity has increased by 400 percent.
The National Kidney Registry is a nonprofit organization that facilitates transplant exchanges. Windows Azure provides a platform-as-a-service (PaaS) environment for organizations to develop applications. It offers automatic OS and service patching, built-in network load balancing and protection from hardware failure.
"More and more transplant centers are using kidney paired exchange transplantation to help their patients in kidney failure, and the computational power required to support the sophisticated matching algorithms that sift through and evaluate all the potential matches has been growing exponentially," Dr. Marc Melcher, transplant surgeon at Stanford University and research director for National Kidney Registry, said in a statement.
Launched on March 9, the registry's Simultaneous Mutually Exclusive Loops and Chains (SMELAC) matching system is an algorithm that allows searchers to perform loop matching, in which one donation leads to reciprocation by someone the recipient knows, and chain matching, in which donations multiply in chains.
The National Kidney Registry announced the migration of SMELAC to Azure on June 5 at the American Transplant Congress in Boston.
The registry used Microsoft's .NET platform to develop SMELAC. Microsoft donated programming and design time to help create the algorithm, Garet Hil, CEO of the National Kidney Registry, told eWEEK.
Although SMELAC requires more computing power than the registry's previous algorithm, it's able to generate more matches, said Hil.
By using Azure, the registry is now able to run multiple match runs simultaneously. "We generally run multiple iterations of the match runs as we zero in on a solution," said Hil.
Before migrating SMELAC to Azure, finding eight donor matches would take 8 hours, Hil explained. On Azure, SMELAC can now process four match runs in 1 hour.
"It increases our time to compute by four times," said Hil. "As the pool gets bigger, that processing window will continue to expand."
The National Kidney Registry is also able to take advantage of Azure's variable computing ability, Wes Anderson, vice president for U.S. public sector services at Microsoft, told eWEEK. This flexibility will allow the registry to grow its computing power exponentially and adjust as needs change rather than being limited to on-premises infrastructure, he said.
"In this world of kidney matching, as our volume goes up, the compute resource to handle simply 100 percent a year could be 100,000 percent a year because the compute requirement goes up exponentially as the pool expands linearly," said Hil.
SMELAC might be able to handle 100,000 times more cycles two years from now in the cloud. "That's where cloud computing and Azure fit in," said Hil. "You've got to be on the platform that can expand the voracious appetite of what's required when we get to a larger pool of incompatible pairs."
Azure is also able to automate the matching process, said Anderson.
"Anyone can virtualize, but oftentimes it's a very manual process, a serial process, and we've automated that with Windows Azure," said Anderson.
In addition, Azure will allow the National Kidney Registry to parallelize its SMELAC software to run multiple processes simultaneously.
On June 7 Microsoft announced updates to Azure to make it more flexible, including the addition of Azure virtual machines. These virtual machines enable infrastructure-as-a-service (IaaS) capabilities for Windows and Linux-based workloads.
Microsoft will provide technical support as the National Kidney Registry grows its searching capabilities, said Hil.