To Queue or Not to Queue: When Queueing is Better Than Timesharing in a Distributed System

We examine the question of whether to employ the first-come-first-served (FCFS) discipline or the processor-sharing (PS) discipline at the hosts in a distributed server system. We are interested in the case in which service times are drawn from a heavy-tailed distribution, and so have very high variability. Traditional wisdom when task sizes are highly variable would prefer the PS discipline, because it allows small tasks to avoid being delayed behind large tasks in a queue. However, we show that system performance can actually be significantly better under FCFS queueing, if each task is assigned to a host based on the task's size. By task assignment, we mean an algorithm that inspects incoming tasks and assigns them to hosts for service. The particular task assignment policy we propose is called SITA-E: Size Interval Task Assignment with Equal Load. Surprisingly, under SITA-E, FCFS queueing typically outperforms the PS discipline by a factor of about two, as measured by mean waiting time and mean slowdown (waiting time of task divided by its service time). We compare the FCFS/SITA-E policy to the processor-sharing case analytically; in addition we compare it to a number of other policies in simulation. We show that the benefits of SITA-E are present even in small-scale distributed systems (four or more hosts). Furthermore, SITA-E is a static policy that does not incorporate feedback knowledge of the state of the hosts, which allows for a simple and scalable implementation.

Task Assignment in a Distributed System: Improving Performance by Unbalancing Load

We consider the problem of task assignment in a distributed system (such as a distributed Web server) in which task sizes are drawn from a heavy-tailed distribution. Many task assignment algorithms are based on the heuristic that balancing the load at the server hosts will result in optimal performance. We show this conventional wisdom is less true when the task size distribution is heavy-tailed (as is the case for Web file sizes). We introduce a new task assignment policy, called Size Interval Task Assignment with Variable Load (SITA-V). SITA-V purposely operates the server hosts at different loads, and directs smaller tasks to the lighter-loaded hosts. The result is that SITA-V provably decreases the mean task slowdown by significant factors (up to 1000 or more) where the more heavy-tailed the workload, the greater the improvement factor. We evaluate the tradeoff between improvement in slowdown and increase in waiting time in a system using SITA-V, and show conditions under which SITA-V represents a particularly appealing policy. We conclude with a discussion of the use of SITA-V in a distributed Web server, and show that it is attractive because it has a simple implementation which requires no communication from the server hosts back to the task router.

Simple Load Balancing for Distributed Hash Tables

Distributed hash tables have recently become a useful building block for a variety of distributed applications. However, current schemes based upon consistent hashing require both considerable implementation complexity and substantial storage overhead to achieve desired load balancing goals. We argue in this paper that these goals can b e achieved more simply and more cost-effectively. First, we suggest the direct application of the "power of two choices" paradigm, whereby an item is stored at the less loaded of two (or more) random alternatives. We then consider how associating a small constant number of hash values with a key can naturally b e extended to support other load balancing methods, including load-stealing or load-shedding schemes, as well as providing natural fault-tolerance mechanisms.

Cluster-Based Optimizations for Distributed Hash Tables

We consider the problem of performing topological optimizations of distributed hash tables. Such hash tables include Chord and Tapestry and are a popular building block for distributed applications. Optimizing topologies over one dimensional hash spaces is particularly difficult as the higher dimensionality of the underlying network makes close fits unlikely. Instead, current schemes are limited to heuristically performing local optimizations finding the best of small random set of peers. We propose a new class of topology optimizations based on the existence of clusters of close overlay members within the underlying network. By constructing additional overlays for each cluster, a significant portion of the search procedure can be performed within the local cluster with a corresponding reduction in the search time. Finally, we discuss the effects of these additional overlays on spatial locality and other load balancing scheme.

Multiplicative Kernels: Object Detection, Segmentation and Pose Estimation

Object detection is challenging when the object class exhibits large within-class variations. In this work, we show that foreground-background classification (detection) and within-class classification of the foreground class (pose estimation) can be jointly learned in a multiplicative form of two kernel functions. One kernel measures similarity for foreground-background classification. The other kernel accounts for latent factors that control within-class variation and implicitly enables feature sharing among foreground training samples. Detector training can be accomplished via standard SVM learning. The resulting detectors are tuned to specific variations in the foreground class. They also serve to evaluate hypotheses of the foreground state. When the foreground parameters are provided in training, the detectors can also produce parameter estimate. When the foreground object masks are provided in training, the detectors can also produce object segmentation. The advantages of our method over past methods are demonstrated on data sets of human hands and vehicles.

An Online Distributed Algorithm for Inferring Policy Routing Configurations

We present an online distributed algorithm, the Causation Logging Algorithm (CLA), in which Autonomous Systems (ASes) in the Internet individually report route oscillations/flaps they experience to a central Internet Routing Registry (IRR). The IRR aggregates these reports and may observe what we call causation chains where each node on the chain caused a route flap at the next node along the chain. A chain may also have a causation cycle. The type of an observed causation chain/cycle allows the IRR to infer the underlying policy routing configuration (i.e., the system of economic relationships and constraints on route/path preferences). Our algorithm is based on a formal policy routing model that captures the propagation dynamics of route flaps under arbitrary changes in topology or path preferences. We derive invariant properties of causation chains/cycles for ASes which conform to economic relationships based on the popular Gao-Rexford model. The Gao-Rexford model is known to be safe in the sense that the system always converges to a stable set of paths under static conditions. Our CLA algorithm recovers the type/property of an observed causation chain of an underlying system and determines whether it conforms to the safe economic Gao-Rexford model. Causes for nonconformity can be diagnosed by comparing the properties of the causation chains with those predicted from different variants of the Gao-Rexford model.

The Sensor Network Workbench: Towards Functional Specification, Verification and Deployment of Constrained Distributed Systems

As the commoditization of sensing, actuation and communication hardware increases, so does the potential for dynamically tasked sense and respond networked systems (i.e., Sensor Networks or SNs) to replace existing disjoint and inflexible special-purpose deployments (closed-circuit security video, anti-theft sensors, etc.). While various solutions have emerged to many individual SN-centric challenges (e.g., power management, communication protocols, role assignment), perhaps the largest remaining obstacle to widespread SN deployment is that those who wish to deploy, utilize, and maintain a programmable Sensor Network lack the programming and systems expertise to do so. The contributions of this thesis centers on the design, development and deployment of the SN Workbench (snBench). snBench embodies an accessible, modular programming platform coupled with a flexible and extensible run-time system that, together, support the entire life-cycle of distributed sensory services. As it is impossible to find a one-size-fits-all programming interface, this work advocates the use of tiered layers of abstraction that enable a variety of high-level, domain specific languages to be compiled to a common (thin-waist) tasking language; this common tasking language is statically verified and can be subsequently re-translated, if needed, for execution on a wide variety of hardware platforms. snBench provides: (1) a common sensory tasking language (Instruction Set Architecture) powerful enough to express complex SN services, yet simple enough to be executed by highly constrained resources with soft, real-time constraints, (2) a prototype high-level language (and corresponding compiler) to illustrate the utility of the common tasking language and the tiered programming approach in this domain, (3) an execution environment and a run-time support infrastructure that abstract a collection of heterogeneous resources into a single virtual Sensor Network, tasked via this common tasking language, and (4) novel formal methods (i.e., static analysis techniques) that verify safety properties and infer implicit resource constraints to facilitate resource allocation for new services. This thesis presents these components in detail, as well as two specific case-studies: the use of snBench to integrate physical and wireless network security, and the use of snBench as the foundation for semester-long student projects in a graduate-level Software Engineering course.

Service Provisioning in Mobile Networks Through Distributed Coordinated Resource Management

The pervasiveness of personal computing platforms offers an unprecedented opportunity to deploy large-scale services that are distributed over wide physical spaces. Two major challenges face the deployment of such services: the often resource-limited nature of these platforms, and the necessity of preserving the autonomy of the owner of these devices. These challenges preclude using centralized control and preclude considering services that are subject to performance guarantees. To that end, this thesis advances a number of new distributed resource management techniques that are shown to be effective in such settings, focusing on two application domains: distributed Field Monitoring Applications (FMAs), and Message Delivery Applications (MDAs). In the context of FMA, this thesis presents two techniques that are well-suited to the fairly limited storage and power resources of autonomously mobile sensor nodes. The first technique relies on amorphous placement of sensory data through the use of novel storage management and sample diffusion techniques. The second approach relies on an information-theoretic framework to optimize local resource management decisions. Both approaches are proactive in that they aim to provide nodes with a view of the monitored field that reflects the characteristics of queries over that field, enabling them to handle more queries locally, and thus reduce communication overheads. Then, this thesis recognizes node mobility as a resource to be leveraged, and in that respect proposes novel mobility coordination techniques for FMAs and MDAs. Assuming that node mobility is governed by a spatio-temporal schedule featuring some slack, this thesis presents novel algorithms of various computational complexities to orchestrate the use of this slack to improve the performance of supported applications. The findings in this thesis, which are supported by analysis and extensive simulations, highlight the importance of two general design principles for distributed systems. First, a-priori knowledge (e.g., about the target phenomena of FMAs and/or the workload of either FMAs or DMAs) could be used effectively for local resource management. Second, judicious leverage and coordination of node mobility could lead to significant performance gains for distributed applications deployed over resource-impoverished infrastructures.

Collocation Games and Their Application to Distributed Resource Management

We introduce Collocation Games as the basis of a general framework for modeling, analyzing, and facilitating the interactions between the various stakeholders in distributed systems in general, and in cloud computing environments in particular. Cloud computing enables fixed-capacity (processing, communication, and storage) resources to be offered by infrastructure providers as commodities for sale at a fixed cost in an open marketplace to independent, rational parties (players) interested in setting up their own applications over the Internet. Virtualization technologies enable the partitioning of such fixed-capacity resources so as to allow each player to dynamically acquire appropriate fractions of the resources for unencumbered use. In such a paradigm, the resource management problem reduces to that of partitioning the entire set of applications (players) into subsets, each of which is assigned to fixed-capacity cloud resources. If the infrastructure and the various applications are under a single administrative domain, this partitioning reduces to an optimization problem whose objective is to minimize the overall deployment cost. In a marketplace, in which the infrastructure provider is interested in maximizing its own profit, and in which each player is interested in minimizing its own cost, it should be evident that a global optimization is precisely the wrong framework. Rather, in this paper we use a game-theoretic framework in which the assignment of players to fixed-capacity resources is the outcome of a strategic "Collocation Game". Although we show that determining the existence of an equilibrium for collocation games in general is NP-hard, we present a number of simplified, practically-motivated variants of the collocation game for which we establish convergence to a Nash Equilibrium, and for which we derive convergence and price of anarchy bounds. In addition to these analytical results, we present an experimental evaluation of implementations of some of these variants for cloud infrastructures consisting of a collection of multidimensional resources of homogeneous or heterogeneous capacities. Experimental results using trace-driven simulations and synthetically generated datasets corroborate our analytical results and also illustrate how collocation games offer a feasible distributed resource management alternative for autonomic/self-organizing systems, in which the adoption of a global optimization approach (centralized or distributed) would be neither practical nor justifiable.

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.

Load Balancing a Cluster of Web Servers using Distributed Packet Rewriting

In this paper, we propose and evaluate an implementation of a prototype scalable web server. The prototype consists of a load-balanced cluster of hosts that collectively accept and service TCP connections. The host IP addresses are advertised using the Round Robin DNS technique, allowing any host to receive requests from any client. Once a client attempts to establish a TCP connection with one of the hosts, a decision is made as to whether or not the connection should be redirected to a different host---namely, the host with the lowest number of established connections. We use the low-overhead Distributed Packet Rewriting (DPR) technique to redirect TCP connections. In our prototype, each host keeps information about connections in hash tables and linked lists. Every time a packet arrives, it is examined to see if it has to be redirected or not. Load information is maintained using periodic broadcasts amongst the cluster hosts.

A Fully Distributed Location Management Scheme for Large PCS

In [previous papers] we presented the design, specification and proof of correctness of a fully distributed location management scheme for PCS networks and argued that fully replicating location information is both appropriate and efficient for small PCS networks. In this paper, we analyze the performance of this scheme. Then, we extend the scheme in a hierarchical environment so as to scale to large PCS networks. Through extensive numerical results, we show the superiority of our scheme compared to the current IS-41 standard.

Type Inference for Variant Object Types

Existing type systems for object calculi are based on invariant subtyping. Subtyping invariance is required for soundness of static typing in the presence of method overrides, but it is often in the way of the expressive power of the type system. Flexibility of static typing can be recovered in different ways: in first-order systems, by the adoption of object types with variance annotations, in second-order systems by resorting to Self types. Type inference is known to be P-complete for first-order systems of finite and recursive object types, and NP-complete for a restricted version of Self types. The complexity of type inference for systems with variance annotations is yet unknown. This paper presents a new object type system based on the notion of Split types, a form of object types where every method is assigned two types, namely, an update type and a select type. The subtyping relation that arises for Split types is variant and, as a result, subtyping can be performed both in width and in depth. The new type system generalizes all the existing first-order type systems for objects, including systems based on variance annotations. Interestingly, the additional expressive power does not affect the complexity of the type inference problem, as we show by presenting an O(n^3) inference algorithm.

Robust Sketching and Aggregation of Distributed Data Streams

The data streaming model provides an attractive framework for one-pass summarization of massive data sets at a single observation point. However, in an environment where multiple data streams arrive at a set of distributed observation points, sketches must be computed remotely and then must be aggregated through a hierarchy before queries may be conducted. As a result, many sketch-based methods for the single stream case do not apply directly, as either the error introduced becomes large, or because the methods assume that the streams are non-overlapping. These limitations hinder the application of these techniques to practical problems in network traffic monitoring and aggregation in sensor networks. To address this, we develop a general framework for evaluating and enabling robust computation of duplicate-sensitive aggregate functions (e.g., SUM and QUANTILE), over data produced by distributed sources. We instantiate our approach by augmenting the Count-Min and Quantile-Digest sketches to apply in this distributed setting, and analyze their performance. We conclude with experimental evaluation to validate our analysis.

Foreground Object Segmentation from Binocular Stereo Video

Moving cameras are needed for a wide range of applications in robotics, vehicle systems, surveillance, etc. However, many foreground object segmentation methods reported in the literature are unsuitable for such settings; these methods assume that the camera is fixed and the background changes slowly, and are inadequate for segmenting objects in video if there is significant motion of the camera or background. To address this shortcoming, a new method for segmenting foreground objects is proposed that utilizes binocular video. The method is demonstrated in the application of tracking and segmenting people in video who are approximately facing the binocular camera rig. Given a stereo image pair, the system first tries to find faces. Starting at each face, the region containing the person is grown by merging regions from an over-segmented color image. The disparity map is used to guide this merging process. The system has been implemented on a consumer-grade PC, and tested on video sequences of people indoors obtained from a moving camera rig. As can be expected, the proposed method works well in situations where other foreground-background segmentation methods typically fail. We believe that this superior performance is partly due to the use of object detection to guide region merging in disparity/color foreground segmentation, and partly due to the use of disparity information available with a binocular rig, in contrast with most previous methods that assumed monocular sequences.