We now have an official GitHub — Have a look at https://github.com/anlab-kaist.

High-performance Network Packet Processors

High-performance Network Platforms based on commodity technologies are only becoming a reality in recent years with the development of high-speed NICs and multi-core platforms. Yet numerous challenges lie ahead to bring down the cost of network platforms and make them more flexible and programmable than today. Heterogeneous coprocessors such as GPUs, Xeon Phis, Tileras, and FGPAs offer a wide range of choices, but complex stacks of software solutions are not always straightforward to take full advantage of. Our SmartPortBox Project aims at building a 1Tbps-grade software router, of which techniques are fully exploitable and applicable not only on routers but also on diverse forms of high-performance networking platforms.

Related Project Sites

Social networking services

How "influence" is "manipulated"?

Online Social Media connect people on earth in intractably many ways and forms in this 21st century. For every form of new media, we need to get used to the new way of online interaction. How effective is the new medium? Why is it effective? How general are the new mechanisms online? What are the consequences and impacts? The goal of our online social media analysis is to discern new forms of online interaction and expound on their intrinsic mechanisms for general applicability.

Related Project Sites

Educational Networking Frameworks

Based on KENS (KAIST Educational Networking System) by Prof. Junewha Song, we have re-designed it as a discrete event simulator with a modular programming interface. We provide a set of hands-on programming assignments on implementation of a TCP/IP stack to KAIST undergraduate students with a reproducible and automated test suite.

Related Project Sites

Future Internet Design

What is wrong with today's Internet and
How do we want to build it again?

Internet, as it grew out of a simple packet-switched networking idea, has lived up way beyond anyone's expectations. Now that it is the network onto which all networking services (phone, fax, radio, TV broadcasting, movies) plus new applications are converging, numerous kludges have been added for a quick fix. What would a computer scientist do given a chance to design a new network? This is the time. Future Internet is a new clean slate approach to draw up a new design for a network that satisfies what today's Internet cannot. We have begun our investigation into fundamental problems that today's Internet cannot address and designs that could turn into real systems

Related Project Sites

Measurement and Infrastructure Building for Emerging Media (VoIP/IPTV/VoD and 3G/3.5G/WiBro/DTN)

Today's Internet carries a myriad of applications and network operators are always playing catch-up with the increase and changes in the traffic. We have conducted an extensive set of measurement experiments to capture the latest trend in the Internet traffic (e.g., UGC video popularity, IPTV over IP-layer multicast, VoIP over WiBro, 3G and 3.5G traffic), and transalted the garnered insight into infrastructure building in order to support the emerging media.

Related Project Sites