Sungwon Ha

The TIMELY adaptive resource management architecture

Mobile computing environments that seek to support communication-intensive applications need to provide sustained end-to-end networking resources to static and mobile flows in the presence of scarce and variable wireless bandwidth, bursty wireless channel error, and user mobility. In order to achieve this goal, we present the TIMELY adaptive resource management architecture and algorithms for resource reservation, advance reservation, and resource adaptation in mobile computing environments. A key feature of our approach is the coordination of adaptation between the different layers of the network in order to solve the problems introduced by scarce and dynamic network resources.

IRMA: a reliable multicast architecture for the Internet

The Illinois Reliable Multicast Architecture (IRMA) is a reliable multicast architecture that guarantees reliable, sequenced, and loosely synchronized delivery of multicast streams. IRMA provides ACK-based reliability, hybrid local server-initiated and sender-initiated loss recovery, and support for end-to-end flow control and congestion control. Unlike most contemporary work, IRMA supports TCP as the reliable multicast transport protocol at the end host without modifications. IRMA has been instantiated in a laboratory testbed, and its performance has been measured in various scenarios. Preliminary performance results show that IRMA is efficient and adaptive to the dynamics of the network.

An adaptive transport protocol for multimedia communication

We propose a novel approach for effectively supporting multimedia packet flows in an Internet environment. The following are the key contributions of the paper: (a) we propose a new transport protocol called HPF that supports multiple interleaved reliable and unreliable data substreams; (b) we decouple the congestion control and reliability mechanisms that are integrated in TCP-this allows us to provide congestion control for unreliable streams as well as reliable streams; and (c) we use priorities within substreams of a packet flow as application defined hints for the link level scheduler to drop packets during congestion. We show through performance measurements in our experimental testbed that our approach can provide effective support for heterogeneous packet flows in the presence of dynamic networking resources.

A scalable router mechanism for load adaptive fair packet dropping

The need for simple, scalable, and incrementally deployable router-level mechanisms to support end-to-end congestion control has been well recognized in the literature. Specifically, mechanisms for early congestion detection and fair packet dropping, such as RED, have been proposed and extensively studied. It has been shown that RED does not scale well with the number of flows, is difficult to tune dynamically, and can degenerate into tail dropping behavior under a variety of conditions. To address these problems while still retaining the robust and simple of RED, we propose a new mechanism called load adaptive fair drop (LAFD), that has no tunable parameters, and adapts effectively to the offered load and the number of flows without maintaining any pre-flow state.

A programmable switch architecture for achieving flexible quality of service

This paper describes a programmable switch architecture that can be dynamically tuned to support flexible quality of service metrics for flows in the Internet. The switch architecture consists of six key components: scheduling, flow aggregation, queueing, packet-level admission control, packet dropping, and feedback. We show that it is possible to achieve a wide range of QoS behavior ranging from strict per-flow rate and delay guarantees to coarse per-class rate assurances by programming the components appropriately.

Performance evaluation of scheduling algorithms in an integrated packet services network environment

This paper studies the performance of a large class of scheduling algorithms, and investigates the interaction between the application and the network to improve performance under congestion. The following key ideas are presented in this paper: (a) we show the performance and scalability trade-offs between providing separation and multiplexing among flows; (b) we show that a bounded buffer FIFO scheduler performs approximately as well as a weighted round robin scheduler with per-flow queues in most practical situations, but requires significantly less overhead in terms of per-flow state; and (c) we show how link layer schedulers can use application-level hints in order to increase the perceived goodness of connections at higher layers.