RAPID: Reliable Adaptive Protocol for Improving Delay And Distributed Computing Programming on .Net
Event details
| Date | 19.06.2015 |
| Hour | 14:00 › 15:00 |
| Location | |
| Category | Conferences - Seminars |
By: Sanjeev Mehrotra – Microsoft Research
Abstract
Low latency networking is becoming critical as an increasing number of cloud applications require interactivity. One way to reduce latency is by placing computation close to the edge. However, this is not always feasible and even then there may be congestion or loss on the network resulting in increased application latency.
In this talk, I will discuss RAPID, a Reliable Adaptive Protocol for Improving Delay which utilizes proactive forward error correction (FEC) and a Universal Rate Control Protocol (URCP) on top of UDP to deliver TCP like semantics for applications, which provides fair and full link utilization, while significantly reducing latency across a wide range of networks. This protocol is currently shipping in several products which require interactivity.
In the second portion of the talk, I will introduce some new work on a suite of tools we are developing for .Net to make it easy to write and deploy distributed systems, including large scale data analytics.
Abstract
Low latency networking is becoming critical as an increasing number of cloud applications require interactivity. One way to reduce latency is by placing computation close to the edge. However, this is not always feasible and even then there may be congestion or loss on the network resulting in increased application latency.
In this talk, I will discuss RAPID, a Reliable Adaptive Protocol for Improving Delay which utilizes proactive forward error correction (FEC) and a Universal Rate Control Protocol (URCP) on top of UDP to deliver TCP like semantics for applications, which provides fair and full link utilization, while significantly reducing latency across a wide range of networks. This protocol is currently shipping in several products which require interactivity.
In the second portion of the talk, I will introduce some new work on a suite of tools we are developing for .Net to make it easy to write and deploy distributed systems, including large scale data analytics.
Practical information
- General public
- Free
- This event is internal
Contact
- Host : Jim Larus