Ken Birman

Ken Birman

See http://www.cs.cornell.edu/ken or follow @KenBirman. My research is on reliable, secure, scalable distributed systems. In ... | Ithaca, New York, United States

Ken Birman Email Addresses

ke****@co****.edu Work Email
ke****@co****.edu Personal Email

Ken Birman Phone Numbers

No phone number available

Ken Birman

Ken Birman Linkedin
Ken Birman Work History
Isis Distributed Systems
Founder and CEO, Chief Scientist
1 January 1988 — 1 January 1996
Cornell University
Professor
1 September 1982 — Present
Ken Birman Current Work Details
Ken Birman's Location
Ithaca, New York, United States
Ken Birman's Expertise
  • See http://www.cs.cornell.edu/ken or follow @KenBirman. My research is on reliable, secure, scalable distributed systems. In the past I've run companies (founded 3 of them, 1 of which became very successful); examples of systems we built include the software that ran the NYSE trading floor for about 15 years, the Swiss Exchange, the French Air Traffic Control System communication platform, and a core component of the US Navy AEGIS warship. More recent work has focused on cloud computing. Right now our Cornell effort is exploring two main ideas: * Cascade, a new platform we've created for speeding up AI by offering cloud platforms a way to run portions of AI and ML inference or classification tasks in secure, private edge settings, and arranging that the code will be physically close to data it depends on. Our work is achieving big speedups in a very transparent way. For example, we have one complicated AI for detecting imminent collisions in traffic intersections. Running on a standard cloud, it needs 2-3 seconds to sense a problem. The identical code, unmodied, runs in milliseconds on Cascade. * Derecho, a solution for accelerating communication on clusters or in clouds by leveraging modern networking technologies like the DPDK protocol stack, and modern networking features like RDMA. Derecho includes ways to perform ultra-fast RPC, multicast, data replication and even has its own set of collective communication primitives (the same as other CCL libraries such as NCCL, Open MPI CCL, etc). We are the world's very fast software stack in all of these areas. For example, many people use RaFT for data replication. Derecho is as much as 150x faster for the same tasks. MPI's CCL is very widely used in AI training. Our DCCL library is 2x faster for tasks like AllReduce. All of this work is open source. You can visit our Derecho repository at https://GitHub.com/Derecho-Project, and can find Cascade at https://GitHub.com/Derecho-Project/cascade
  • N. Rama Rao Professor at Cornell University and Distributed Systems Researcher
Ken Birman's Current Industry
Isis Distributed Systems
Ken Birman's Prior Industry
Cornell University | Isis Distributed Systems

Ken Birman Education

College Logo
Camden County College
Camden County College Associate's degree, Business Administration and Management, General

Not the Ken Birman you were looking for?

Find accurate emails & phone numbers for over 700M professionals.

Search

FAQ About Ken Birman

  • Ken Birman is located at Ithaca, New York, United States

  • Ken Birman is a See http://www.cs.cornell.edu/ken or follow @KenBirman. My research is on reliable, secure, scalable distributed systems. In the past I've run companies (founded 3 of them, 1 of which became very successful); examples of systems we built include the software that ran the NYSE trading floor for about 15 years, the Swiss Exchange, the French Air Traffic Control System communication platform, and a core component of the US Navy AEGIS warship. More recent work has focused on cloud computing. Right now our Cornell effort is exploring two main ideas: * Cascade, a new platform we've created for speeding up AI by offering cloud platforms a way to run portions of AI and ML inference or classification tasks in secure, private edge settings, and arranging that the code will be physically close to data it depends on. Our work is achieving big speedups in a very transparent way. For example, we have one complicated AI for detecting imminent collisions in traffic intersections. Running on a standard cloud, it needs 2-3 seconds to sense a problem. The identical code, unmodied, runs in milliseconds on Cascade. * Derecho, a solution for accelerating communication on clusters or in clouds by leveraging modern networking technologies like the DPDK protocol stack, and modern networking features like RDMA. Derecho includes ways to perform ultra-fast RPC, multicast, data replication and even has its own set of collective communication primitives (the same as other CCL libraries such as NCCL, Open MPI CCL, etc). We are the world's very fast software stack in all of these areas. For example, many people use RaFT for data replication. Derecho is as much as 150x faster for the same tasks. MPI's CCL is very widely used in AI training. Our DCCL library is 2x faster for tasks like AllReduce. All of this work is open source. You can visit our Derecho repository at https://GitHub.com/Derecho-Project, and can find Cascade at https://GitHub.com/Derecho-Project/cascade Isis Distributed Systems

  • Ken Birman currently works in the Isis Distributed Systems

  • Ken Birman specializes in See http://www.cs.cornell.edu/ken or follow @KenBirman. My research is on reliable, secure, scalable distributed systems. In the past I've run companies (founded 3 of them, 1 of which became very successful); examples of systems we built include the software that ran the NYSE trading floor for about 15 years, the Swiss Exchange, the French Air Traffic Control System communication platform, and a core component of the US Navy AEGIS warship. More recent work has focused on cloud computing. Right now our Cornell effort is exploring two main ideas: * Cascade, a new platform we've created for speeding up AI by offering cloud platforms a way to run portions of AI and ML inference or classification tasks in secure, private edge settings, and arranging that the code will be physically close to data it depends on. Our work is achieving big speedups in a very transparent way. For example, we have one complicated AI for detecting imminent collisions in traffic intersections. Running on a standard cloud, it needs 2-3 seconds to sense a problem. The identical code, unmodied, runs in milliseconds on Cascade. * Derecho, a solution for accelerating communication on clusters or in clouds by leveraging modern networking technologies like the DPDK protocol stack, and modern networking features like RDMA. Derecho includes ways to perform ultra-fast RPC, multicast, data replication and even has its own set of collective communication primitives (the same as other CCL libraries such as NCCL, Open MPI CCL, etc). We are the world's very fast software stack in all of these areas. For example, many people use RaFT for data replication. Derecho is as much as 150x faster for the same tasks. MPI's CCL is very widely used in AI training. Our DCCL library is 2x faster for tasks like AllReduce. All of this work is open source. You can visit our Derecho repository at https://GitHub.com/Derecho-Project, and can find Cascade at https://GitHub.com/Derecho-Project/cascade

  • Ken Birman speaks the following languages: No languages available.

  • Ken Birman has prior experience in the following industries: Isis Distributed Systems

  • Ken Birman has the following professional experience: Isis Distributed Systems , Cornell University .

  • You can view Ken Birman’s LinkedIn profile at LinkedIn URL.

  • Ken Birman’s work email addresses are [email protected] .

  • Ken Birman works at Isis Distributed Systems