Results 1  10
of
16
Pathlet routing
 In Proc. SIGCOMM Workshop on Hot Topics in Networking
, 2008
"... We present a new routing protocol, pathlet routing, in which networks advertise fragments of paths, called pathlets, that sources concatenate into endtoend source routes. Intuitively, the pathlet is a highly flexible building block, capturing policy constraints as well as enabling an exponentially ..."
Abstract

Cited by 37 (11 self)
 Add to MetaCart
We present a new routing protocol, pathlet routing, in which networks advertise fragments of paths, called pathlets, that sources concatenate into endtoend source routes. Intuitively, the pathlet is a highly flexible building block, capturing policy constraints as well as enabling an exponentially large number of path choices. In particular, we show that pathlet routing can emulate the policies of BGP, source routing, and several recent multipath proposals. This flexibility lets us address two major challenges for Internet routing: scalability and sourcecontrolled routing. When a router’s routing policy has only “local ” constraints, it can be represented using a small number of pathlets, leading to very small forwarding tables and many choices of routes for senders. Crucially, pathlet routing does not impose a global requirement on what style of policy is used, but rather allows multiple styles to coexist. The protocol thus supports complex routing policies while enabling and incentivizing the adoption of policies that yield small forwarding plane state and a high degree of path choice.
Online Decision Problems with Large Strategy Sets
, 2005
"... In an online decision problem, an algorithm performs a sequence of trials, each of which involves selecting one element from a fixed set of alternatives (the “strategy set”) whose costs vary over time. After T trials, the combined cost of the algorithm’s choices is compared with that of the single s ..."
Abstract

Cited by 24 (2 self)
 Add to MetaCart
In an online decision problem, an algorithm performs a sequence of trials, each of which involves selecting one element from a fixed set of alternatives (the “strategy set”) whose costs vary over time. After T trials, the combined cost of the algorithm’s choices is compared with that of the single strategy whose combined cost is minimum. Their difference is called regret, and one seeks algorithms which are efficient in that their regret is sublinear in T and polynomial in the problem size. We study an important class of online decision problems called generalized multiarmed bandit problems. In the past such problems have found applications in areas as diverse as statistics, computer science, economic theory, and medical decisionmaking. Most existing algorithms were efficient only in the case of a small (i.e. polynomialsized) strategy set. We extend the theory by supplying nontrivial algorithms and lower bounds for cases in which the strategy set is much larger (exponential or infinite) and
The online shortest path problem under partial monitoring
 JOURNAL OF MACHINE LEARNING RESEARCH
, 2007
"... The online shortest path problem is considered under partial monitoring scenarios. At each round, a decision maker has to choose a path between two distinguished vertices of a weighted directed acyclic graph whose edge weights can change in an arbitrary (adversarial) way such that the loss of the ..."
Abstract

Cited by 18 (5 self)
 Add to MetaCart
The online shortest path problem is considered under partial monitoring scenarios. At each round, a decision maker has to choose a path between two distinguished vertices of a weighted directed acyclic graph whose edge weights can change in an arbitrary (adversarial) way such that the loss of the chosen path (defined as the sum of the weights of its composing edges) be small. In the multiarmed bandit setting, after choosing a path, the decision maker learns only the weights of those edges that belong to the chosen path. For this scenario, an algorithm is given whose average cumulative loss in n rounds exceeds that of the best path, matched offline to the entire sequence of the edge weights, by a quantity that is proportional to 1 / √n and depends only polynomially on the number of edges of the graph. The algorithm can be implemented with linear complexity in the number of rounds n and in the number of edges. This result improves earlier banditalgorithms which have performance bounds that either depend exponentially on the number of edges or converge to zero at a slower rate than O(1 / √n). An extension to the socalled label efficient setting is also given, where the decision maker is informed about the weight of the chosen path only with probability ɛ < 1. Applications to routing in packet switched networks along with simulation results are also presented.
HighProbability Regret Bounds for Bandit Online Linear Optimization
"... We present a modification of the algorithm of Dani et al. [8] for the online linear optimization problem in the bandit setting, which with high probability has regret at most O ∗ ( √ T) against an adaptive adversary. This improves on the previous algorithm [8] whose regret is bounded in expectatio ..."
Abstract

Cited by 16 (0 self)
 Add to MetaCart
We present a modification of the algorithm of Dani et al. [8] for the online linear optimization problem in the bandit setting, which with high probability has regret at most O ∗ ( √ T) against an adaptive adversary. This improves on the previous algorithm [8] whose regret is bounded in expectation against an oblivious adversary. We obtain the same dependence on the dimension (n 3/2) as that exhibited by Dani et al. The results of this paper rest firmly on those of [8] and the remarkable technique of Auer et al. [2] for obtaining highprobability bounds via optimistic estimates. This paper answers an open question: it eliminates the gap between the highprobability bounds obtained in the fullinformation vs bandit settings. 1
From optimization to regret minimization and back again
 In Proc. Third Workshop on Tackling Computer System Problems with Machine Learning Techniques (SysML08
, 2008
"... Internet routing is mostly based on static information— it’s dynamicity is limited to reacting to changes in topology. Adaptive performancebased routing decisions would not only improve the performance itself of the Internet but also its security and availability. However, previous approaches for m ..."
Abstract

Cited by 2 (1 self)
 Add to MetaCart
Internet routing is mostly based on static information— it’s dynamicity is limited to reacting to changes in topology. Adaptive performancebased routing decisions would not only improve the performance itself of the Internet but also its security and availability. However, previous approaches for making Internet routing adaptive based on optimizing networkwide objectives are not suited for an environment in which autonomous and possibly malicious entities interact. In this paper, we propose a different framework for adaptive routing decisions based on regretminimizing online learning algorithms. These algorithms, as applied to routing, are appealing because adopters can independently improve their own performance while being robust to adversarial behavior. However, in contrast to approaches based on optimization theory that provide guarantees from the outset about networkwide behavior, the networkwide behavior if online learning algorithms were to interact with each other is less understood. In this paper, we study this interaction in a realistic Internet environment, and find that the outcome is a stable state and that the optimality gap with respect to the networkwide optimum is small. Our findings suggest that online learning may be a suitable framework for adaptive routing decisions in the Internet. 1
Throughput Optimal OnLine Algorithms for Advanced Resource Reservation in Ultra HighSpeed Networks
"... Abstract—Advanced channel reservation is emerging as an important feature of ultra highspeed networks requiring the transfer of large files. Applications include scientific data transfers and database backup. In this paper, we present two new, online algorithms for advanced reservation, called Batc ..."
Abstract

Cited by 1 (1 self)
 Add to MetaCart
Abstract—Advanced channel reservation is emerging as an important feature of ultra highspeed networks requiring the transfer of large files. Applications include scientific data transfers and database backup. In this paper, we present two new, online algorithms for advanced reservation, called BatchAll and BatchLim, that are guaranteed to achieve optimal throughput performance, based on multicommodity flow arguments. Both algorithms are shown to have polynomialtime complexity and provable bounds on the maximum delay for 1 + ε bandwidth augmented networks. The BatchLim algorithm returns the completion time of a connection immediately as a request is placed, but at the expense of a slightly looser competitive ratio than that of BatchAll. We also present a simple approach that limits the number of parallel paths used by the algorithms while provably bounding the maximum reduction factor in the transmission throughput. We show that, although the number of different paths can be exponentially large, the actual number of paths needed to approximate the flow is quite small and proportional to the number of edges in the network. Simulations for a number of topologies show that, in practice, 3 to 5 parallel paths are sufficient to achieve close to optimal performance. The performance of the competitive algorithms are also compared to a greedy benchmark, both through analysis and simulation. I.
ABSTRACT Online Collaborative Filtering with Nearly Optimal Dynamic Regret
"... We consider a model for sequential online decisionmaking by many diverse agents. On each day, each agent makes a decision, and pays a penalty if it is a mistake. Obviously, it would be good for agents to avoid repeating the same mistakes made by other agents; however, difficulty may arise when some ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
We consider a model for sequential online decisionmaking by many diverse agents. On each day, each agent makes a decision, and pays a penalty if it is a mistake. Obviously, it would be good for agents to avoid repeating the same mistakes made by other agents; however, difficulty may arise when some agents disagree over what constitutes a mistake, perhaps maliciously. As a metric of success for this problem, we consider dynamic regret, i.e., regret versus the offline optimal sequence of decisions. Previous regret bounds usually use the much weaker notion of static regret, i.e., regret versus the best single decision in hindsight. We assume there is a set of “honest ” players whose valuations for the decisions at each time step are identical. No assumptions are made about the remaining players, and the algorithm assumes no information about which are the honest players. We present an algorithm for this setting whose expected dynamic regret per honest player is optimal up to a multiplicative constant and an additive polylogarithmic term, assuming the number of options is bounded. Categories and Subject Descriptors I.2.11 [Artificial Intelligence]: Distributed artificial intelligence—Multiagent systems; G.3 [Probability and Statistics]:
ThroughputCompetitive Advance Reservation with Bounded Path Dispersion
, 2008
"... In response to the high throughput needs of grid and cloud computing applications, several production networks have recently started to support advance reservation of dedicated circuits. An important open problem within this context is to devise advance reservation algorithms that can provide prova ..."
Abstract
 Add to MetaCart
In response to the high throughput needs of grid and cloud computing applications, several production networks have recently started to support advance reservation of dedicated circuits. An important open problem within this context is to devise advance reservation algorithms that can provide provable throughput performance guarantees, independently of the specific network topology and arrival pattern of reservation requests. In this paper, we first show that the throughput performance of greedy approaches, which return the earliest possible completion time for each incoming request, can be arbitrarily worse than optimal. Next, we introduce two new online, polynomialtime algorithms for advance reservation, called BatchAll and BatchLim. Both algorithms are shown to be throughputoptimal through the derivation of delay bounds for 1 + ε bandwidth augmented networks. The BatchLim algorithm has the advantage of returning the completion time of a connection immediately as a request is placed, but at the expense of looser delay performance than BatchAll. We then propose a simple approach that limits path dispersion, i.e., the number of parallel paths used by the algorithms, while provably bounding the maximum reduction factor in the transmission throughput. We prove that the number of paths needed to approximate any flow is quite small and never exceeds the total number of edges in the network. Through simulation for various topologies and traffic parameters, we show that the proposed algorithms achieve reasonable delay performance, even at request arrival rates close to capacity bounds, and that three to five parallel paths are sufficient to achieve nearoptimal performance.
1 ThroughputCompetitive Advance Reservation with Bounded Path Dispersion
"... Abstract—In response to the high throughput needs of grid and cloud computing applications, several production networks have recently started to support advance reservation of dedicated circuits. An important open problem within this context is to devise advance reservation algorithms that can provi ..."
Abstract
 Add to MetaCart
Abstract—In response to the high throughput needs of grid and cloud computing applications, several production networks have recently started to support advance reservation of dedicated circuits. An important open problem within this context is to devise advance reservation algorithms that can provide provable throughput performance guarantees, independently of the specific network topology and arrival pattern of reservation requests. In this paper, we first show that the throughput performance of greedy approaches, which return the earliest possible completion time for each incoming request, can be arbitrarily worse than optimal. Next, we introduce two new online, polynomialtime algorithms for advance reservation, calledBatchAll andBatchLim. Both algorithms are shown to be throughputoptimal through the derivation of bounds on the maximum delay for 1+ε bandwidth augmented networks. TheBatchLim algorithm has the advantage of returning the completion time of a connection immediately as a request is placed, but at the expense of looser delay performance than BatchAll. We then propose a simple approach that limits path dispersion, i.e., the number of parallel paths used by the algorithms, while provably bounding the maximum reduction factor in the transmission throughput. We prove that, although the number of different paths can be exponentially large, the actual number of paths needed to approximate any flow is quite small and proportional to the total number of edges in the network. Through simulation for various topologies and traffic parameters, we show that the proposed algorithms achieve reasonable average delay performance, even at network loads close to capacity bounds, and that three to five parallel paths are sufficient to achieve performance close to optimal.
Hungarian Academy of Sciences
"... The online shortest path problem is considered under various models of partial monitoring. Given a weighted directed acyclic graph whose edge weights can change in an arbitrary (adversarial) way, a decision maker has to choose in each round of a game a path between two distinguished vertices such t ..."
Abstract
 Add to MetaCart
The online shortest path problem is considered under various models of partial monitoring. Given a weighted directed acyclic graph whose edge weights can change in an arbitrary (adversarial) way, a decision maker has to choose in each round of a game a path between two distinguished vertices such that the loss of the chosen path (defined as the sum of the weights of its composing edges) be as small as possible. In a setting generalizing the multiarmed bandit problem, after choosing a path, the decision maker learns only the weights of those edges that belong to the chosen path. For this problem, an algorithm is given whose average cumulative loss in n rounds exceeds that of the best path, matched offline to the entire sequence of the edge weights, by a quantity that is proportional to 1 / √ n and depends only polynomially on the number of edges of the graph. The algorithm can be implemented with complexity that is linear in the number of rounds n (i.e., the average complexity per round is constant) and in the number of edges. An extension to the socalled label efficient setting is also given, in which the decision maker is informed about the weights of the edges corresponding to the chosen path at a total of m ≪ n time instances. Another extension is shown where the decision maker competes against a timevarying path, a generalization of the