Keon Jang | KAIST - Academia.edu (original) (raw)
Papers by Keon Jang
Today massively-parallel processors are becoming increasingly popular and cost-effective. Future ... more Today massively-parallel processors are becoming increasingly popular and cost-effective. Future high performance networking systems design should harness the full computation power of such massively-parallel processors effecitvely for various types of workloads, beyond simple multi-threading on homogeneous fat-core processors. Software routers are no exception.
Abstract The Click modular router has been one of the most popular software router platforms for ... more Abstract The Click modular router has been one of the most popular software router platforms for rapid prototyping and new protocol development. Unfortunately, its internal architecture has not caught up with recent hardware advancements, and the performance remains sub-optimal in high-speed networks despite its benefit of flexible module composition. In this work, we identify the performance bottlenecks of the existing Click router and extend it to scale with modern computer systems.
Abstract With the rapid growth of data centers, minimizing the queueing delay at network switches... more Abstract With the rapid growth of data centers, minimizing the queueing delay at network switches has been one of the key challenges. In this work, we analyze the shortcomings of the current TCP algorithm when used in data center networks, and we propose to use latency-based congestion detection and rate-based transfer to achieve ultra-low queueing delay in data centers.
The current Internet has only two name spaces: IP addresses and domain names [1]. Domain names ar... more The current Internet has only two name spaces: IP addresses and domain names [1]. Domain names are rather aliases of IP addresses than identifier, which leaves only one name spaace in the Internet. The lack of name space leads to an overloaded semantics (locator, host identifier, and routing identifier) on the IP address [2]. One problem with the overloading is that it is hard to support mobility.[2] Since TCP connections are established using IP and port, connection is not sustained when node change its IP address.
ACM SIGCOMM Computer …, Jan 1, 2010
Proceedings of the 9th …, Jan 1, 2008
Proceedings of the 1st …, Jan 1, 2009
Proceedings of the …, Jan 1, 2009
ROADS, October, Jan 1, 2009
Proc. of Conference …, Jan 1, 2008
INFOCOM, 2010 …, Jan 1, 2010
CFI'08: Proceedings of the 3rd International …, Jan 1, 2008
ACM SIGCOMM Computer …, Jan 1, 2010
… Department, KAIST, Tech …, Jan 1, 2009
Local and Metropolitan …, Jan 1, 2010
Abstract Commodity-hardware technology has advanced in great leaps in terms of CPU, memory, and I... more Abstract Commodity-hardware technology has advanced in great leaps in terms of CPU, memory, and I/O bus speeds. Benefiting from the hardware innovation, recent software routers on commodity PC now report about 10 Gbps in packet routing. In this paper we map out expected hurdles and projected speed-ups to reach 100 Gbps in packet routing on a single commodity PC. With careful measurements, we identify two notable bottlenecks for our goal: CPU cycles and I/O bandwidth. For the former, we propose reducing per-packet ...
Proceedings of the 2008 …, Jan 1, 2008
Today massively-parallel processors are becoming increasingly popular and cost-effective. Future ... more Today massively-parallel processors are becoming increasingly popular and cost-effective. Future high performance networking systems design should harness the full computation power of such massively-parallel processors effecitvely for various types of workloads, beyond simple multi-threading on homogeneous fat-core processors. Software routers are no exception.
Abstract The Click modular router has been one of the most popular software router platforms for ... more Abstract The Click modular router has been one of the most popular software router platforms for rapid prototyping and new protocol development. Unfortunately, its internal architecture has not caught up with recent hardware advancements, and the performance remains sub-optimal in high-speed networks despite its benefit of flexible module composition. In this work, we identify the performance bottlenecks of the existing Click router and extend it to scale with modern computer systems.
Abstract With the rapid growth of data centers, minimizing the queueing delay at network switches... more Abstract With the rapid growth of data centers, minimizing the queueing delay at network switches has been one of the key challenges. In this work, we analyze the shortcomings of the current TCP algorithm when used in data center networks, and we propose to use latency-based congestion detection and rate-based transfer to achieve ultra-low queueing delay in data centers.
The current Internet has only two name spaces: IP addresses and domain names [1]. Domain names ar... more The current Internet has only two name spaces: IP addresses and domain names [1]. Domain names are rather aliases of IP addresses than identifier, which leaves only one name spaace in the Internet. The lack of name space leads to an overloaded semantics (locator, host identifier, and routing identifier) on the IP address [2]. One problem with the overloading is that it is hard to support mobility.[2] Since TCP connections are established using IP and port, connection is not sustained when node change its IP address.
ACM SIGCOMM Computer …, Jan 1, 2010
Proceedings of the 9th …, Jan 1, 2008
Proceedings of the 1st …, Jan 1, 2009
Proceedings of the …, Jan 1, 2009
ROADS, October, Jan 1, 2009
Proc. of Conference …, Jan 1, 2008
INFOCOM, 2010 …, Jan 1, 2010
CFI'08: Proceedings of the 3rd International …, Jan 1, 2008
ACM SIGCOMM Computer …, Jan 1, 2010
… Department, KAIST, Tech …, Jan 1, 2009
Local and Metropolitan …, Jan 1, 2010
Abstract Commodity-hardware technology has advanced in great leaps in terms of CPU, memory, and I... more Abstract Commodity-hardware technology has advanced in great leaps in terms of CPU, memory, and I/O bus speeds. Benefiting from the hardware innovation, recent software routers on commodity PC now report about 10 Gbps in packet routing. In this paper we map out expected hurdles and projected speed-ups to reach 100 Gbps in packet routing on a single commodity PC. With careful measurements, we identify two notable bottlenecks for our goal: CPU cycles and I/O bandwidth. For the former, we propose reducing per-packet ...
Proceedings of the 2008 …, Jan 1, 2008