A Two-step Statistical Approach for Inferring Network Traffic Demands (Revises Technical Report BUCS-2003-003)

Accurate knowledge of traffic demands in a communication network enables or enhances a variety of traffic engineering and network management tasks of paramount importance for operational networks. Directly measuring a complete set of these demands is prohibitively expensive because of the huge amounts of data that must be collected and the performance impact that such measurements would impose on the regular behavior of the network. As a consequence, we must rely on statistical techniques to produce estimates of actual traffic demands from partial information. The performance of such techniques is however limited due to their reliance on limited information and the high amount of computations they incur, which limits their convergence behavior. In this paper we study a two-step approach for inferring network traffic demands. First we elaborate and evaluate a modeling approach for generating good starting points to be fed to iterative statistical inference techniques. We call these starting points informed priors since they are obtained using actual network information such as packet traces and SNMP link counts. Second we provide a very fast variant of the EM algorithm which extends its computation range, increasing its accuracy and decreasing its dependence on the quality of the starting point. Finally, we evaluate and compare alternative mechanisms for generating starting points and the convergence characteristics of our EM algorithm against a recently proposed Weighted Least Squares approach.

Quadsim Version 2.1 Student Manual

Quadsim is an intermediate code simulator. It allows you to "run" programs that your compiler generates in intermediate code format. Its user interface is similar to most debuggers in that you can step through your program, instruction by instruction, set breakpoints, examine variable values, and so on. The intermediate code format used by Quadsim is that described in [Aho 86]. If your compiler generates intermediate code in this format, you will be able to take intermediate-code files generated by your compiler, load them into the simulator, and watch them "run." You are provided with functions that hide the internal representation of intermediate code. You can use these functions within your compiler to generate intermediate code files that can be read by the simulator. Quadsim was inspired and greatly influenced by [Aho 86]. The material in chapter 8 (Intermediate Code Generation) of [Aho 86] should be considered background material for users of Quadsim.

Mapping Parallel Iterative Algorithms onto Workstation Networks

For communication-intensive parallel applications, the maximum degree of concurrency achievable is limited by the communication throughput made available by the network. In previous work [HPS94], we showed experimentally that the performance of certain parallel applications running on a workstation network can be improved significantly if a congestion control protocol is used to enhance network performance. In this paper, we characterize and analyze the communication requirements of a large class of supercomputing applications that fall under the category of fixed-point problems, amenable to solution by parallel iterative methods. This results in a set of interface and architectural features sufficient for the efficient implementation of the applications over a large-scale distributed system. In particular, we propose a direct link between the application and network layer, supporting congestion control actions at both ends. This in turn enhances the system's responsiveness to network congestion, improving performance. Measurements are given showing the efficacy of our scheme to support large-scale parallel computations.

On the Convergence of Statistical Techniques for Inferring Network Traffic Demands

Accurate knowledge of traffic demands in a communication network enables or enhances a variety of traffic engineering and network management tasks of paramount importance for operational networks. Directly measuring a complete set of these demands is prohibitively expensive because of the huge amounts of data that must be collected and the performance impact that such measurements would impose on the regular behavior of the network. As a consequence, we must rely on statistical techniques to produce estimates of actual traffic demands from partial information. The performance of such techniques is however limited due to their reliance on limited information and the high amount of computations they incur, which limits their convergence behavior. In this paper we study strategies to improve the convergence of a powerful statistical technique based on an Expectation-Maximization iterative algorithm. First we analyze modeling approaches to generating starting points. We call these starting points informed priors since they are obtained using actual network information such as packet traces and SNMP link counts. Second we provide a very fast variant of the EM algorithm which extends its computation range, increasing its accuracy and decreasing its dependence on the quality of the starting point. Finally, we study the convergence characteristics of our EM algorithm and compare it against a recently proposed Weighted Least Squares approach.

Bounds on the Power of Constant-Depth Quantum Circuits

We show that if a language is recognized within certain error bounds by constant-depth quantum circuits over a finite family of gates, then it is computable in (classical) polynomial time. In particular, our results imply EQNC^0 ⊆ P, where EQNC^0 is the constant-depth analog of the class EQP. On the other hand, we adapt and extend ideas of Terhal and DiVincenzo [?] to show that, for any family

Demand-based Document Dissemination for the World-Wide Web

We analyzed the logs of our departmental HTTP server http://cs-www.bu.edu as well as the logs of the more popular Rolling Stones HTTP server http://www.stones.com. These servers have very different purposes; the former caters primarily to local clients, whereas the latter caters exclusively to remote clients all over the world. In both cases, our analysis showed that remote HTTP accesses were confined to a very small subset of documents. Using a validated analytical model of server popularity and file access profiles, we show that by disseminating the most popular documents on servers (proxies) closer to the clients, network traffic could be reduced considerably, while server loads are balanced. We argue that this process could be generalized so as to provide for an automated demand-based duplication of documents. We believe that such server-based information dissemination protocols will be more effective at reducing both network bandwidth and document retrieval times than client-based caching protocols [2].

Visible Volume: a Robust Measure for Protein Structure Characterization

We propose a new characterization of protein structure based on the natural tetrahedral geometry of the β carbon and a new geometric measure of structural similarity, called visible volume. In our model, the side-chains are replaced by an ideal tetrahedron, the orientation of which is fixed with respect to the backbone and corresponds to the preferred rotamer directions. Visible volume is a measure of the non-occluded empty space surrounding each residue position after the side-chains have been removed. It is a robust, parameter-free, locally-computed quantity that accounts for many of the spatial constraints that are of relevance to the corresponding position in the native structure. When computing visible volume, we ignore the nature of both the residue observed at each site and the ones surrounding it. We focus instead on the space that, together, these residues could occupy. By doing so, we are able to quantify a new kind of invariance beyond the apparent variations in protein families, namely, the conservation of the physical space available at structurally equivalent positions for side-chain packing. Corresponding positions in native structures are likely to be of interest in protein structure prediction, protein design, and homology modeling. Visible volume is related to the degree of exposure of a residue position and to the actual rotamers in native proteins. In this article, we discuss the properties of this new measure, namely, its robustness with respect to both crystallographic uncertainties and naturally occurring variations in atomic coordinates, and the remarkable fact that it is essentially independent of the choice of the parameters used in calculating it. We also show how visible volume can be used to align protein structures, to identify structurally equivalent positions that are conserved in a family of proteins, and to single out positions in a protein that are likely to be of biological interest. These properties qualify visible volume as a powerful tool in a variety of applications, from the detailed analysis of protein structure to homology modeling, protein structural alignment, and the definition of better scoring functions for threading purposes.

BRITE: Universal Topology Generation from a User's Perspective

Effective engineering of the Internet is predicated upon a detailed understanding of issues such as the large-scale structure of its underlying physical topology, the manner in which it evolves over time, and the way in which its constituent components contribute to its overall function. Unfortunately, developing a deep understanding of these issues has proven to be a challenging task, since it in turn involves solving difficult problems such as mapping the actual topology, characterizing it, and developing models that capture its emergent behavior. Consequently, even though there are a number of topology models, it is an open question as to how representative the topologies they generate are of the actual Internet. Our goal is to produce a topology generation framework which improves the state of the art and is based on design principles which include representativeness, inclusiveness, and interoperability. Representativeness leads to synthetic topologies that accurately reflect many aspects of the actual Internet topology (e.g. hierarchical structure, degree distribution, etc.). Inclusiveness combines the strengths of as many generation models as possible in a single generation tool. Interoperability provides interfaces to widely-used simulation and visualization applications such as ns and SSF. We call such a tool a universal topology generator. In this paper we discuss the design, implementation and usage of the BRITE universal topology generation tool that we have built. We also describe the BRITE Analysis Engine, BRIANA, which is an independent piece of software designed and built upon BRITE design goals of flexibility and extensibility. The purpose of BRIANA is to act as a repository of analysis routines along with a user–friendly interface that allows its use on different topology formats.

Deanonymizing Users of the SafeWeb Anonymizing Service

The SafeWeb anonymizing system has been lauded by the press and loved by its users; self-described as "the most widely used online privacy service in the world," it served over 3,000,000 page views per day at its peak. SafeWeb was designed to defeat content blocking by firewalls and to defeat Web server attempts to identify users, all without degrading Web site behavior or requiring users to install specialized software. In this article we describe how these fundamentally incompatible requirements were realized in SafeWeb's architecture, resulting in spectacular failure modes under simple JavaScript attacks. These exploits allow adversaries to turn SafeWeb into a weapon against its users, inflicting more damage on them than would have been possible if they had never relied on SafeWeb technology. By bringing these problems to light, we hope to remind readers of the chasm that continues to separate popular and technical notions of security.

Parameter Sensitive Detectors

Object detection can be challenging when the object class exhibits large variations. One commonly-used strategy is to first partition the space of possible object variations and then train separate classifiers for each portion. However, with continuous spaces the partitions tend to be arbitrary since there are no natural boundaries (for example, consider the continuous range of human body poses). In this paper, a new formulation is proposed, where the detectors themselves are associated with continuous parameters, and reside in a parameterized function space. There are two advantages of this strategy. First, a-priori partitioning of the parameter space is not needed; the detectors themselves are in a parameterized space. Second, the underlying parameters for object variations can be learned from training data in an unsupervised manner. In profile face detection experiments, at a fixed false alarm number of 90, our method attains a detection rate of 75% vs. 70% for the method of Viola-Jones. In hand shape detection, at a false positive rate of 0.1%, our method achieves a detection rate of 99.5% vs. 98% for partition based methods. In pedestrian detection, our method reduces the miss detection rate by a factor of three at a false positive rate of 1%, compared with the method of Dalal-Triggs.

An Information Theoretic Framework for Field Monitoring Using Autonomously Mobile Sensors

We consider a mobile sensor network monitoring a spatio-temporal field. Given limited cache sizes at the sensor nodes, the goal is to develop a distributed cache management algorithm to efficiently answer queries with a known probability distribution over the spatial dimension. First, we propose a novel distributed information theoretic approach in which the nodes locally update their caches based on full knowledge of the space-time distribution of the monitored phenomenon. At each time instant, local decisions are made at the mobile nodes concerning which samples to keep and whether or not a new sample should be acquired at the current location. These decisions account for minimizing an entropic utility function that captures the average amount of uncertainty in queries given the probability distribution of query locations. Second, we propose a different correlation-based technique, which only requires knowledge of the second-order statistics, thus relaxing the stringent constraint of having a priori knowledge of the query distribution, while significantly reducing the computational overhead. It is shown that the proposed approaches considerably improve the average field estimation error by maintaining efficient cache content. It is further shown that the correlation-based technique is robust to model mismatch in case of imperfect knowledge of the underlying generative correlation structure.

Angels: In Network Support for Minimum Distribution Time in P2P Overlays

This paper proposes the use of in-network caches (which we call Angels) to reduce the Minimum Distribution Time (MDT) of a file from a seeder – a node that possesses the file – to a set of leechers – nodes who are interested in downloading the file. An Angel is not a leecher in the sense that it is not interested in receiving the entire file, but rather it is interested in minimizing the MDT to all leechers, and as such uses its storage and up/down-link capacity to cache and forward parts of the file to other peers. We extend the analytical results by Kumar and Ross [1] to account for the presence of angels by deriving a new lower bound for the MDT. We show that this newly derived lower bound is tight by proposing a distribution strategy under assumptions of a fluid model. We present a GroupTree heuristic that addresses the impracticalities of the fluid model. We evaluate our designs through simulations that show that our Group-Tree heuristic outperforms other heuristics, that it scales well with the increase of the number of leechers, and that it closely approaches the optimal theoretical bounds.

Distributed Packet Rewriting and its Application to Scalable Server Architectures

To construct high performance Web servers, system builders are increasingly turning to distributed designs. An important challenge that arises in distributed Web servers is the need to direct incoming connections to individual hosts. Previous methods for connection routing have employed a centralized node which handles all incoming requests. In contrast, we propose a distributed approach, called Distributed Packet Rewriting (DPR), in which all hosts of the distributed system participate in connection routing. We argue that this approach promises better scalability and fault-tolerance than the centralized approach. We describe our implementation of four variants of DPR and compare their performance. We show that DPR provides performance comparable to centralized alternatives, measured in terms of throughput and delay under the SPECweb96 benchmark. Finally, we argue that DPR is particularly attractive both for small scale systems and for systems following the emerging trend toward increasingly intelligent I/O subsystems.

Connection Scheduling in Web Servers

Under high loads, a Web server may be servicing many hundreds of connections concurrently. In traditional Web servers, the question of the order in which concurrent connections are serviced has been left to the operating system. In this paper we ask whether servers might provide better service by using non-traditional service ordering. In particular, for the case when a Web server is serving static files, we examine the costs and benefits of a policy that gives preferential service to short connections. We start by assessing the scheduling behavior of a commonly used server (Apache running on Linux) with respect to connection size and show that it does not appear to provide preferential service to short connections. We then examine the potential performance improvements of a policy that does favor short connections (shortest-connection-first). We show that mean response time can be improved by factors of four or five under shortest-connection-first, as compared to an (Apache-like) size-independent policy. Finally we assess the costs of shortest-connection-first scheduling in terms of unfairness (i.e., the degree to which long connections suffer). We show that under shortest-connection-first scheduling, long connections pay very little penalty. This surprising result can be understood as a consequence of heavy-tailed Web server workloads, in which most connections are small, but most server load is due to the few large connections. We support this explanation using analysis.

On the Complexity of Quantum ACC

For any q > 1, let MOD_q be a quantum gate that determines if the number of 1's in the input is divisible by q. We show that for any q,t > 1, MOD_q is equivalent to MOD_t (up to constant depth). Based on the case q=2, Moore has shown that quantum analogs of AC^(0), ACC[q], and ACC, denoted QAC^(0)_wf, QACC[2], QACC respectively, define the same class of operators, leaving q > 2 as an open question. Our result resolves this question, implying that QAC^(0)_wf = QACC[q] = QACC for all q. We also prove the first upper bounds for QACC in terms of related language classes. We define classes of languages EQACC, NQACC (both for arbitrary complex amplitudes) and BQACC (for rational number amplitudes) and show that they are all contained in TC^(0). To do this, we show that a TC^(0) circuit can keep track of the amplitudes of the state resulting from the application of a QACC operator using a constant width polynomial size tensor sum. In order to accomplish this, we also show that TC^(0) can perform iterated addition and multiplication in certain field extensions.