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.

Performance Evaluation of Two-Shadow Speculative Concurrency Control

Speculative Concurrency Control (SCC) [Best92a] is a new concurrency control approach especially suited for real-time database applications. It relies on the use of redundancy to ensure that serializable schedules are discovered and adopted as early as possible, thus increasing the likelihood of the timely commitment of transactions with strict timing constraints. In [Best92b], SCC-nS, a generic algorithm that characterizes a family of SCC-based algorithms was described, and its correctness established by showing that it only admits serializable histories. In this paper, we evaluate the performance of the Two-Shadow SCC algorithm (SCC-2S), a member of the SCC-nS family, which is notable for its minimal use of redundancy. In particular, we show that SCC-2S (as a representative of SCC-based algorithms) provides significant performance gains over the widely used Optimistic Concurrency Control with Broadcast Commit (OCC-BC), under a variety of operating conditions and workloads.

Proceedings of Sixth International Workshop on Unification

Swiss National Science Foundation; Austrian Federal Ministry of Science and Research; Deutsche Forschungsgemeinschaft (SFB 314); Christ Church, Oxford; Oxford University Computing Laboratory

Mermera: Non-Coherent Distributed Shared Memory for Parallel Computing

The proliferation of inexpensive workstations and networks has prompted several researchers to use such distributed systems for parallel computing. Attempts have been made to offer a shared-memory programming model on such distributed memory computers. Most systems provide a shared-memory that is coherent in that all processes that use it agree on the order of all memory events. This dissertation explores the possibility of a significant improvement in the performance of some applications when they use non-coherent memory. First, a new formal model to describe existing non-coherent memories is developed. I use this model to prove that certain problems can be solved using asynchronous iterative algorithms on shared-memory in which the coherence constraints are substantially relaxed. In the course of the development of the model I discovered a new type of non-coherent behavior called Local Consistency. Second, a programming model, Mermera, is proposed. It provides programmers with a choice of hierarchically related non-coherent behaviors along with one coherent behavior. Thus, one can trade-off the ease of programming with coherent memory for improved performance with non-coherent memory. As an example, I present a program to solve a linear system of equations using an asynchronous iterative algorithm. This program uses all the behaviors offered by Mermera. Third, I describe the implementation of Mermera on a BBN Butterfly TC2000 and on a network of workstations. The performance of a version of the equation solving program that uses all the behaviors of Mermera is compared with that of a version that uses coherent behavior only. For a system of 1000 equations the former exhibits at least a 5-fold improvement in convergence time over the latter. The version using coherent behavior only does not benefit from employing more than one workstation to solve the problem while the program using non-coherent behavior continues to achieve improved performance as the number of workstations is increased from 1 to 6. This measurement corroborates our belief that non-coherent shared memory can be a performance boon for some applications.

An Implementation of Mermera: A Shared Memory System that Mixes Coherence with Non-coherence

Coherent shared memory is a convenient, but inefficient, method of inter-process communication for parallel programs. By contrast, message passing can be less convenient, but more efficient. To get the benefits of both models, several non-coherent memory behaviors have recently been proposed in the literature. We present an implementation of Mermera, a shared memory system that supports both coherent and non-coherent behaviors in a manner that enables programmers to mix multiple behaviors in the same program[HS93]. A programmer can debug a Mermera program using coherent memory, and then improve its performance by selectively reducing the level of coherence in the parts that are critical to performance. Mermera permits a trade-off of coherence for performance. We analyze this trade-off through measurements of our implementation, and by an example that illustrates the style of programming needed to exploit non-coherence. We find that, even on a small network of workstations, the performance advantage of non-coherence is compelling. Raw non-coherent memory operations perform 20-40~times better than non-coherent memory operations. An example application program is shown to run 5-11~times faster when permitted to exploit non-coherence. We conclude by commenting on our use of the Isis Toolkit of multicast protocols in implementing Mermera.

On Learning Counting Functions With Queries

We investigate the problem of learning disjunctions of counting functions, which are general cases of parity and modulo functions, with equivalence and membership queries. We prove that, for any prime number p, the class of disjunctions of integer-weighted counting functions with modulus p over the domain Znq (or Zn) for any given integer q ≥ 2 is polynomial time learnable using at most n + 1 equivalence queries, where the hypotheses issued by the learner are disjunctions of at most n counting functions with weights from Zp. The result is obtained through learning linear systems over an arbitrary field. In general a counting function may have a composite modulus. We prove that, for any given integer q ≥ 2, over the domain Zn2, the class of read-once disjunctions of Boolean-weighted counting functions with modulus q is polynomial time learnable with only one equivalence query, and the class of disjunctions of log log n Boolean-weighted counting functions with modulus q is polynomial time learnable. Finally, we present an algorithm for learning graph-based counting functions.

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.

A Hybrid GLR Algorithm for Parsing with Epsilon Grammars

We give a hybrid algorithm for parsing epsilon grammars based on Tomita's non-ϵ-grammar parsing algorithm ([Tom86]) and Nozohoor-Farshi's ϵ-grammar recognition algorithm ([NF91]). The hybrid parser handles the same set of grammars handled by Nozohoor-Farshi's recognizer. The algorithm's details and an example of its use are given. We also discuss the deployment of the hybrid algorithm within a GB parser, and the reason an ϵ grammar parser is needed in our GB parser.

Structure Sharing and Parallelization in a GB Parser

By utilizing structure sharing among its parse trees, a GB parser can increase its efficiency dramatically. Using a GB parser which has as its phrase structure recovery component an implementation of Tomita's algorithm (as described in [Tom86]), we investigate how a GB parser can preserve the structure sharing output by Tomita's algorithm. In this report, we discuss the implications of using Tomita's algorithm in GB parsing, and we give some details of the structuresharing parser currently under construction. We also discuss a method of parallelizing a GB parser, and relate it to the existing literature on parallel GB parsing. Our approach to preserving sharing within a shared-packed forest is applicable not only to GB parsing, but anytime we want to preserve structure sharing in a parse forest in the presence of features.

Adding Polymorphic Abstraction to ML (Detailed Abstract)

The ML programming language restricts type polymorphism to occur only in the "let-in" construct and requires every occurrence of a formal parameter of a function (a lambda abstraction) to have the same type. Milner in 1978 refers to this restriction (which was adopted to help ML achieve automatic type inference) as a serious limitation. We show that this restriction can be relaxed enough to allow universal polymorphic abstraction without losing automatic type inference. This extension is equivalent to the rank-2 fragment of system F. We precisely characterize the additional program phrases (lambda terms) that can be typed with this extension and we describe typing anomalies both before and after the extension. We discuss how macros may be used to gain some of the power of rank-3 types without losing automatic type inference. We also discuss user-interface problems in how to inform the programmer of the possible types a program phrase may have.

Building Responsive Systems from Physically-correct Specifications

Predictability - the ability to foretell that an implementation will not violate a set of specified reliability and timeliness requirements - is a crucial, highly desirable property of responsive embedded systems. This paper overviews a development methodology for responsive systems, which enhances predictability by eliminating potential hazards resulting from physically-unsound specifications. The backbone of our methodology is the Time-constrained Reactive Automaton (TRA) formalism, which adopts a fundamental notion of space and time that restricts expressiveness in a way that allows the specification of only reactive, spontaneous, and causal computation. Using the TRA model, unrealistic systems - possessing properties such as clairvoyance, caprice, in finite capacity, or perfect timing - cannot even be specified. We argue that this "ounce of prevention" at the specification level is likely to spare a lot of time and energy in the development cycle of responsive systems - not to mention the elimination of potential hazards that would have gone, otherwise, unnoticed. The TRA model is presented to system developers through the CLEOPATRA programming language. CLEOPATRA features a C-like imperative syntax for the description of computation, which makes it easier to incorporate in applications already using C. It is event-driven, and thus appropriate for embedded process control applications. It is object-oriented and compositional, thus advocating modularity and reusability. CLEOPATRA is semantically sound; its objects can be transformed, mechanically and unambiguously, into formal TRA automata for verification purposes, which can be pursued using model-checking or theorem proving techniques. Since 1989, an ancestor of CLEOPATRA has been in use as a specification and simulation language for embedded time-critical robotic processes.

A Minimal GB Parser

We describe a GB parser implemented along the lines of those written by Fong [4] and Dorr [2]. The phrase structure recovery component is an implementation of Tomita's generalized LR parsing algorithm (described in [10]), with recursive control flow (similar to Fong's implementation). The major principles implemented are government, binding, bounding, trace theory, case theory, θ-theory, and barriers. The particular version of GB theory we use is that described by Haegeman [5]. The parser is minimal in the sense that it implements the major principles needed in a GB parser, and has fairly good coverage of linguistically interesting portions of the English language.

Multi-version Speculative Concurrency Control with Delayed Commit

This paper presents an algorithm which extends the relatively new notion of speculative concurrency control by delaying the commitment of transactions, thus allowing other conflicting transactions to continue execution and commit rather than restart. This algorithm propagates uncommitted data to other outstanding transactions thus allowing more speculative schedules to be considered. The algorithm is shown always to find a serializable schedule, and to avoid cascading aborts. Like speculative concurrency control, it considers strictly more schedules than traditional concurrency control algorithms. Further work is needed to determine which of these speculative methods performs better on actual transaction loads.

How Good are Genetic Algorithms at Finding Large Cliques: An Experimental Study

This paper investigates the power of genetic algorithms at solving the MAX-CLIQUE problem. We measure the performance of a standard genetic algorithm on an elementary set of problem instances consisting of embedded cliques in random graphs. We indicate the need for improvement, and introduce a new genetic algorithm, the multi-phase annealed GA, which exhibits superior performance on the same problem set. As we scale up the problem size and test on \hard" benchmark instances, we notice a degraded performance in the algorithm caused by premature convergence to local minima. To alleviate this problem, a sequence of modi cations are implemented ranging from changes in input representation to systematic local search. The most recent version, called union GA, incorporates the features of union cross-over, greedy replacement, and diversity enhancement. It shows a marked speed-up in the number of iterations required to find a given solution, as well as some improvement in the clique size found. We discuss issues related to the SIMD implementation of the genetic algorithms on a Thinking Machines CM-5, which was necessitated by the intrinsically high time complexity (O(n3)) of the serial algorithm for computing one iteration. Our preliminary conclusions are: (1) a genetic algorithm needs to be heavily customized to work "well" for the clique problem; (2) a GA is computationally very expensive, and its use is only recommended if it is known to find larger cliques than other algorithms; (3) although our customization e ort is bringing forth continued improvements, there is no clear evidence, at this time, that a GA will have better success in circumventing local minima.