The data broadcast problem is to nd a schedule for broadcasting a given set of messages over multiple channels. The goal is to minimize the cost of the broadcast plus the expected response time to clients who periodically and probabilistically tune in to wait for particular messages. The problem models disseminating data to clients in asymmetric communication environments, where there is a much larger capacity from the information source to the clients than in the reverse direction. Examples include satellites, cable TV, internet broadcast, and mobile phones. Such environments favor the \push-based" model where the server broadcasts (pushes) its information on the communication medium and multiple clients simultaneously retrieve the speci c information of individual interest. This sort of environment motivates the study of \broadcast disks" in Information Systems [1; 7]. In this paper we present the rst polynomial-time approximation scheme for the data broadcast problem for the cas...

Abstract not found

We study the problem of scheduling infinitely ¢ often jobs, each with an associated demand probability, under the constraint that each job must be scheduled with a fixed period. That is, the number of time units between two consecutive occurrences of each job is constant (we assume that time is slotted and that each job can be scheduled in a single time-slot). The goal is to minimize the average time a random arriving client waits until its desired job is executed. This problem is a variant of the broadcast disks problem: the perfect periodicity allows clients to know exactly when their job is scheduled, rather than “busy waiting,” thus saving energy. The problem is known to be NP-hard. The best known polynomial algorithm to date guarantees average waiting time of at ¦ §©¨�����������������¢� � most, ¨��© � where is the optimal waiting time. In this paper, we develop a tree-based methodology for periodic scheduling, and using new general results, we derive algorithms with better bounds. A key quantity in our �������� � �� � ������������ � � methodology is. We compare the cost of a solution provided by our algorithms to the cost of a solution to a relaxed continuous (nonintegral) version of the problem. Our asymptotic tree-based algorithm guarantees cost of ��������� at most times the cost of the relaxed problem; on the other hand, we prove that the cost of any integral solution is bounded from below by the cost of the continuous �������� � � solution times. We also provide three other tree-based algorithms with cost bounded by the cost of the continuous solution ���� � times ���������������� � ,,

We consider a variant of the classic multi-armed bandit problem (MAB), which we call FEEDBACK MAB, where the reward obtained by playing each of n independent arms varies according to an underlying on/off Markov process with known parameters. The evolution of the Markov chain happens irrespective of whether the arm is played, and furthermore, the exact state of the Markov chain is only revealed to the player when the arm is played and the reward observed. At most one arm (or in general, M arms) can be played any time step. The goal is to design a policy for playing the arms in order to maximize the infinite horizon time average expected reward. This problem is an instance of a Partially Observable Markov Decision Process (POMDP), and a special case of the notoriously intractable “restless bandit ” problem. Unlike the stochastic MAB problem, the FEEDBACK MAB problem does not admit to greedy index-based optimal policies. The state of the system at any time step encodes the beliefs about the states of different arms, and the policy decisions change these beliefs – this aspect complicates the design and analysis of simple algorithms. We design a constant factor approximation to the FEEDBACK MAB problem by solving and rounding a natural LP relaxation to this problem. As far as we are aware, this is the first approximation algorithm for a POMDP problem. 1

Advances in wireless and optical communication, as well as in Internet multicast protocols, make broadcast and multicast methods an effective solution to disseminate data. In particular, repetitive server-initiated broadcast is an effective technique in wireless systems and is a scalable solution to relieve Internet hot spots. A critical issue for the performance of multicast data dissemination is the multicast schedule. Previous work focused on a model where each data item is requested by clients with a certain probability that is independent of past accesses. In this paper, we consider the more complex scenario where a client accesses pages in blocks (e.g., a HTML file and all its embedded images), thereby introducing dependencies in the pattern of accesses to data. We present a sequence of heuristics that exploit page access dependencies. We measured the resulting client-perceived delay on multiple Web server traces, and observed an average speed-up over previous methods ranging from 8% to 91%. We conclude that scheduling for multi-item requests is a critical factor for the performance of repetitive broadcast.

In satellite and wireless networks and in advanced traffic information systems in which the up-link bandwidth is very limited, a server broadcasts data files in a round-robin manner. The data files are provided by different providers and are accessed by many clients. The providers are independent and therefore files may share information. The clients who access these files may have different patterns of access. Some clients may wish to access more than one file at a time in any order, some clients may access one file out of of several files, and some clients may wish to access a second file only after accessing another file. The goal of the server is to order the files in a way that minimizes the access time of the clients given some a-priori knowledge of their access patterns. This paper introduces a clients-providers-servers model that better represents certain environments than the traditional clients-servers model. Then, we show that a random order of the data files performs well, independent of the specific access pattern. Our main technical contribution is de-randomizing the procedure that is based on selecting a random order. The resulting algorithm is a polynomial-time deterministic algorithm that finds an order with the same performance bounds as those of the random order.

Unicast connections lead to performance and scalability problems when a large client population attempts to access the same data. Broadcast push and broadcast disk technology address the problem by broadcasting data items from a server to a large number of clients. Broadcast disk performance depends mainly on caching strategies at the client site and on how the broadcast is scheduled at the server site. An on-line broadcast disk paging strategy makes caching decisions without knowing access probabilities. In this paper, we subject on-line paging algorithms to extensive empirical investigation. The Gray algorithm [25] always outperformed other on-line strategies on both synthetic and Web traces. Moreover, caching limited the skewness needed from a broadcast schedule, and led to favor efficient caching algorithms over refined scheduling strategies when the cache was not small. Prior to this paper, no work had empirically investigated on-line paging algorithms and their relation with serv...

The first main goal of this paper is to present Sketch-it!, a framework aiming to facilitate development and experimental evaluation of new scheduling algorithms. It comprises many helpful data-structures, a graphical interface with several components and a library with implementations of selected scheduling algorithms. Every scheduling problem covered by the classification-scheme originally proposed by Graham et al. [22] can easily be integrated into the framework. One of the more recent enhancements of this scheme, the so called broadcast scheduling problem, was chosen for an extensive case study of Sketch-it!, yielding very interesting experimental results that represent the second main contribution of this paper. In broadcast scheduling many clients listen to a high bandwidth channel on which a server can transmit documents of a given set. Over time the clients request certain documents. In the pull-based setting each client has access to a slow bandwidth channel whereon it notifies the server about its requests. In the push-based setting no such channel exists. Instead it is assumed that requests for certain documents arrive randomly with probabilities known to the server. The goal in both settings is to generate broadcast schedules for these documents which minimize the average time a client has to wait until a request is answered. We conduct experiments with several algorithms on generated data. We distinguish scenarios for which a slow feedback channel is very advantageous, and others where its benefits are negligible, answering the question posed in the title.

Abstract — The broadcast disk provides a way to distribute data to many clients simultaneously. A central server fixes a set of data and a schedule for sending it, and then repeatedly sends the data according to the schedule. Clients listen for data until it is broadcast. We look at the problem of scheduling for two separate channels, where each can have a different broadcast schedule. Our metric for measuring schedule performance is expected delivery time (EDT), the expected value of the total elapsed time between when a client starts listening for data and when the client is completely finished receiving the data. We fix the first channel with a schedule that is optimal for an average case, and look at how to schedule for the second channel. We show two interesting results for sending two items over two channels. The first is that all schedules with equal portions of the two items in the second channel have the same EDT. The second is that for a situation that is symmetric in the two items the optimal schedule is asymmetric with respect to these items. I.

Abstract not found