Baek-Young Choi

Analysis of point-to-point packet delay in an operational network

In this paper we perform a detailed analysis of point-to-point packet delay in an operational tier-1 network. The point-to-point delay is the time between a packet entering a router in one PoP (an ingress point) and its leaving a router in another PoP (an egress point). It measures the one-way delay experienced by packets from an ingress point to an egress point across an ISPs network and provides the most basic information regarding the delay performance of the ISPs network. Using packet traces captured in the operational network, we obtain precise point-to-point packet delay measurements and analyze the various factors affecting them. Through a simple, step-by-step, systematic methodology and careful data analysis, we identify the major network factors that contribute to point-to-point packet delay and characterize their effect on the network delay performance. Our findings are: 1) delay distributions vary greatly in shape, depending on the path and link utilization; 2) after constant factors dependent only on the path and packet size are removed, the 99th percentile variable delay remains under 1 ms over several hops and under link utilization below 90% on a bottleneck; 3) a very small number of packets experience very large delay in short bursts

Adaptive random sampling for load change detection

Timely detection of changes in traffic load is critical for initiating appropriate traffic engineering mechanisms. Accurate measurement of traffic is essential since the efficacy of change detection depends on the accuracy of traffic estimation. However, precise traffic measurement involves inspecting every packet traversing a link, resulting in significant overhead, particularly on high speed links. Sampling techniques for traffic load estimation are proposed as a way to limit the measurement overhead. In this paper, we address the problem of bounding sampling error within a pre-specified tolerance level and propose an adaptive random sampling technique that determines the minimum sampling probability adaptively according to traffic dynamics. Using real network traffic traces, we show that the proposed adaptive random sampling technique indeed produces the desired accuracy, while also yielding significant reduction in the amount of traffic samples. We also investigate the impact of sampling errors on the performance of load change detection.

Probabilistic approach to switched Ethernet for real-time control applications

Switched Ethernet combines Ethernet and ATM switching technology. The authors propose a probability based admission control algorithm over switched Ethernet to provide real time capability for control applications. The main contribution of the paper is providing an efficient means of estimating the queuing possibilities and the corresponding queuing delays inside the switch for periodic control system application. The approach enables switched Ethernet admission control without sophisticated packet scheduling in switches or middleware on hosts. The algorithm is analyzed and implemented to verify its feasibility.

Observations on Cisco sampled NetFlow

Traffic monitoring is an important first step for network management and traffic engineering. With high-speed Internet backbone links, efficient and effective packet sampling is not only desirable, but also increasingly becoming a necessity. The Sampled NetFlow [10] is Cisco routers traffic measurement functionality with static packet sampling for high speed links. Since the utility of sampling depends on the accuracy and economy of measurement, it is important to understand sampling error and measurement overhead. In this paper, we first discuss fundamental limitations of sampling techniques used in the Sampled NetFlow. We assess the accuracy of the Sampled NetFlow by comparing its output with complete packet traces [8] from an operational router. We also show the overheads involved in the Sampled NetFlow. We find that Sampled NetFlow performs correctly without incurring dramatic overhead during our experiments. However, a care should be taken in its use, since the overhead is linearly proportional to the number of flow records.

Practical delay monitoring for ISPs

Point-to-point delay is an important network performance measure as well as a key parameter in SLAs. We study how to measure and report delay in a concise and meaningful way for an ISP, and how to monitor it efficiently. We analyze various measurement intervals and potential metric definitions. We find that reporting high quantiles (between 0.95 and 0.99)every 10-30 minutes as the most effective way to summarize the delay in an ISP. We then propose an active probing scheme to estimate a high quantile with bounded error. We show that only a small number of probes are sufficient to provide an accurate estimate. We validate the proposed delay monitoring technique on real data collected on the Sprint IP backbone network.

Outage Analysis of a University Campus Network

Understanding outage and failure characteristics of a network is important to assess the availability of the network, determine failure source for trouble-shooting, and identify weak areas for network availability improvement. However, there has been virtually no failure measurement and analysis on access networks. In this paper, we carry out an in-depth outage and failure analysis of a university campus network using a rich set of both node outage and link failure data. We investigate the aspects of spatial and temporal localities of failures and outages, the relation of link failure and node outage, and the impact of the hierarchical and redundant network design on outage. We find most of link failure events are not caused by node failures; frequent link up-down events may not lead to the corresponding nodes outage; for access layer switches that connect to end hosts, their link up-down events exhibit periodic patterns.

Energy efficient virtual network embedding for green data centers using data center topology and future migration

We formulate energy efficient virtual network embedding that incorporates energy costs of operation and migration for nodes and links.We prove the NP-hardness of the problem and develop a heuristic algorithm to minimize the energy consumption.We consider a practical intra-DC architecture to further improve energy efficiency.We conduct extensive evaluations and comparisons with existing algorithms to show that the proposed algorithm substantially saves energy consumption and allows high acceptance ratios. With the rapid proliferation of data centers, their energy consumption and greenhouse gas emissions have significantly increased. Some efforts have been made to control and lower energy consumption of data centers such as, proportional energy consuming hardware, dynamic provisioning, and virtualization machine techniques. However, it is still common that many servers and network resources are often underutilized, and idle servers spend a large portion of their peak power consumption.We first build a novel model of virtual network embedding in order to minimize energy usage in data centers for both computing and network resources by taking practical factors into consideration. Due to the NP-hardness of the proposed model, we develop a heuristic algorithm for virtual network scheduling and mapping. In doing so, we specifically take the expected energy consumption at different times, virtual network operation and future migration costs, and a data center architecture into consideration. Our extensive evaluation results show that our algorithm could reduce energy consumption up to 40% and take up to a 57% higher number of virtual network requests over other existing virtual mapping schemes.

Light Weight Anti-Phishing with User Whitelisting in a Web Browser

Phishing often causes more financial damages to enterprises and users than virus and worms. Proposed techniques so far mostly are to provide protection for the banks or other enterprise corporations and/or require changes in Web servers. Regular home users are left without protection against phishing, other than the traditional rudimentary email filtering. Furthermore, security toolbars on browsers have been shown to be not effective. In this paper, we propose a light-weight user white list technique to provide protection for home users. We argue that for home users, white listing is advantageous over blacklisting, a popular technology. We show it is effective and efficient while imposing little burden on users and requiring no modification of servers.

Protocol independent multicast group aggregation scheme for the global area multicast

IP multicast is an important enabling service for the current and future Internet. With the explosive growth of the Internet, a challenging issue facing IP multicast is scalability, in particular, the problem of multicast forwarding state and control explosion. In this paper, we propose a new methodology to address the multicast scalability problem for backbone domains-multicast tunneling with branch filtering (MTBF). This multicast group aggregation scheme is designed on top of the inter-domain protocol architecture such as MASC/BGMF, and is independent of any underlying intra-domain multicast protocols. It aggregates multicast groups by constructing bolder router (BR)-based multicast routing trees and forwards data by using an encapsulation technique called multicast tunneling (MT). The feasibility and performance of our scheme is demonstrated through analysis and simulations.

NQAR: network quality aware routing in error-prone wireless sensor networks

We propose a network quality aware routing (NQAR) mechanism to provide an enabling method of the delay-sensitive data delivery over error-prone wireless sensor networks. Unlike the existing routing methods that select routes with the shortest arrival latency or the minimum hop count, the proposed scheme adaptively selects the route based on the network qualities including link errors and collisions with minimum additional complexity. It is designed to avoid the paths with potential noise and collision that may cause many non-deterministic backoffs and retransmissions. We propose a generic framework to select a minimum cost route that takes the packet loss rate and collision history into account. NQAR uses a data centric approach to estimate a single-hop delay based on processing time, propagation delay, packet loss rate, number of backoffs, and the retransmission timeout between two neighboring nodes. This enables a source node to choose the shortest expected end-to-end delay path to send a delay-sensitive data. The experiment results show that NQAR reduces the end-to-end transfer delay up to approximately 50% in comparison with the latency-based directed diffusion and the hop count-based directed diffusion under the error-prone network environments. Moreover, NQAR shows better performance than those routing methods in terms of jitter, reachability, and network lifetime.