Abstract: The management of Grid resources requires scheduling of both computation and communication tasks at various levels. In this study, we consider the two constituent sub-problems of Grid scheduling, namely: (i) the scheduling of computation tasks to processing resources and (ii) the routing and scheduling of the data movement in a Grid network. Regarding computation tasks, we examine two typical online task scheduling algorithms that employ advance reservations and perform full network simulation experiments to measure their performance when implemented in a centralized or distributed manner. Similarly, for communication tasks, we compare two routing and data scheduling algorithms that are implemented in a centralized or a distributed manner. We examine the effect network propagation delay has on the performance of these algorithms. Our simulation results indicate that a distributed architecture with an exhaustive resource utilization update strategy yields better average end-to-end delay performance than a centralized architecture.
Abstract: We present an improved upper bound on the competitiveness of the online colouring algorithm First-Fit in disk graphs, which
are graphs representing overlaps of disks on the plane. We also show that this bound is best possible for deterministic online
colouring algorithms that do not use the disk representation of the input graph. We also present a related new lower bound for unit
disk graphs.
Abstract: An ever growing emphasis is put nowadays in developing personalized journey planning and renewable mobility services in smart cities. These services combine means of scheduled-based public transport and electric vehicles or bikes, using crowdsourcing techniques for collecting real-time traffic information and for assessing the recommended routes. The goal is to develop an information system that will allow the fast, real-time computation of best routes.
The main challenges in developing such an information system are both technological and algorithmic. The technological challenge concerns the collection, storage, management, and updating of a huge volume of transport data that are usually time-dependent, and the provision (through these data) of personalized renewable mobility services in smartphones. This challenge is typically confronted by creating a cloud infrastructure that on the one hand will support the storage, management, and updating of data, while on the other hand it will handle the necessary data feed to the smartphone applications for providing the users with the requested best routes.
The algorithmic challenge concerns the development of innovative algorithms for the efficient provision of journey planning services in smartphones, based on data they will receive from the cloud infrastructure. These services guarantee the computation of realistic and useful best routes, as well as the updating of the precomputed (route and timetable) information, in case of delays of scheduled public transport vehicles, so that the users can online update their routes to destination. The goal is to develop an algorithmic basis for supporting modern renewable mobility services (information systems), such as "mobility on demand'' (where the next leg of a journey is decided in real-time) and "door-to-door'' personalized mobility, in urban scheduled-based public transport environments. Scheduled-based public transport information systems should not only compute in real-time end-user queries requesting best routes, but also to update the timetable information in case of delays.
The core algorithmic issues of mobility and journey planning (regarding the computation of optimal routes under certain criteria) in scheduled-based public transport systems concern the efficient solution of the fundamental earlier arrival (EA) problem (compute a journey from station S to station T minimizing the overall traveling time required to complete the journey), the minimum number of
transfers (MNT) problem (compute a journey from station S to station T minimizing the number of times a passenger is required to change vehicle), and the efficient updating of timetable information system in case of vehicle delays. The EA and MNT problems have been extensively studied in the literature under two main approaches: the array-based modeling (where the timetable is represented as an array) and the graph-based modeling (where the timetable is represented as a graph). Experimental results have shown so far that the array-based approaches are faster in terms of query time than graph-based ones, as they are able to better exploit data locality and do not rely on priority queues. On the other hand, the array-based approaches have not been theoretically or experimentally studied as far as the efficient updating of timetable information, in case of delays, is concerned.
In this thesis, new graph-based models are being developed that solve efficiently the aforementioned fundamental algorithmic mobility problems in urban scheduled-based public transport information systems, along with a mobile application (journey planner) running on Android-based smartphones that includes a service for the evaluation of the recommended routes by the users. In particular:
(a) An extensive comparative evaluation was conducted on graph-based dynamic models that represent big data volumes regarding their suitability for representing timetable information. The study confirmed that the realistic time-expanded model is the most suitable for representing timetable information.
(b) Two new graph-based models have been developed for representing timetable information (in a timetable information system), the reduced time-expanded model and the dynamic timetable model (DTM), both of which are more space-efficient with respect to the realistic time-expanded model. For both of the new models, new efficient algorithms were developed for fast answering of EA and MNT queries, as well as for updating the timetable information representation in case of delays.
(c)An experimental evaluation was conducted with the new graph-based models and their associated query and update algorithms on a set of 14 real-world scheduled-based transportation systems, including the metropolitan areas of Berlin, Athens, London, Rome, and Madrid. The experimental results showed that the query algorithms of the reduced time-expanded model are superior to those of the DTM model, while the reverse is true regarding the update algorithms. In addition, the experimental study showed that the query algorithms of the new graph-based models compete favorably with those of the best array-based models.
(d) A mobile, cloud-based, journey planner (information system) was developed whose core algorithmic engine builds upon the new graph-based models. The mobile application is accompanied by a service that allows the users to assess the recommended journeys. The journey planner demonstrates the practicality of the new graph-based models and their associated query and update algorithms.
Abstract: We study the problem of scheduling permanent jobs on un-
related machines when the objective is to minimize the Lp
norm of the machine loads. The problem is known as load
balancing under the Lp norm. We present an improved up-
per bound for the greedy algorithm through simple analy-
sis; this bound is also shown to be best possible within the
class of deterministic onlinealgorithms for the problem. We
also address the question whether randomization helps on-
line load balancing under Lp norms on unrelated machines;
this is a challenging question which is open for more than a
decade even for the L2 norm. We provide a positive answer
to this question by presenting the ¯rst randomized onlinealgorithms which outperform deterministic ones under any
(integral) Lp norm for p = 2; :::; 137. Our algorithms es-
sentially compute in an online manner a fractional solution
to the problem and use the fractional values to make ran-
dom choices. The local optimization criterion used at each
step is novel and rather counterintuitive: the values of the
fractional variables for each job correspond to °ows at an ap-
proximate Wardrop equilibrium for an appropriately de¯ned
non-atomic congestion game. As corollaries of our analysis
and by exploiting the relation between the Lp norm and the
makespan of machine loads, we obtain new competitive algo-
rithms for online makespan minimization, making progress
in another longstanding open problem.
Abstract: We study the problem of scheduling permanent jobs on un-
related machines when the objective is to minimize the Lp
norm of the machine loads. The problem is known as load
balancing under the Lp norm. We present an improved up-
per bound for the greedy algorithm through simple analy-
sis; this bound is also shown to be best possible within the
class of deterministic onlinealgorithms for the problem. We
also address the question whether randomization helps on-
line load balancing under Lp norms on unrelated machines;
this is a challenging question which is open for more than a
decade even for the L2 norm. We provide a positive answer
to this question by presenting the ¯rst randomized onlinealgorithms which outperform deterministic ones under any
(integral) Lp norm for p = 2; :::; 137. Our algorithms es-
sentially compute in an online manner a fractional solution
to the problem and use the fractional values to make ran-
dom choices. The local optimization criterion used at each
step is novel and rather counterintuitive: the values of the
fractional variables for each job correspond to °ows at an ap-
proximate Wardrop equilibrium for an appropriately de¯ned
non-atomic congestion game. As corollaries of our analysis
and by exploiting the relation between the Lp norm and the
makespan of machine loads, we obtain new competitive algo-
rithms for online makespan minimization, making progress
in another longstanding open problem.
Abstract: We address an important communication issue arising in
wireless cellular networks that utilize frequency division
multiplexing (FDM) technology. In such networks, many
users within the same geographical region (cell) can communicate
simultaneously with other users of the network
using distinct frequencies. The spectrum of the available
frequencies is limited; thus, efficient solutions to the call
controlproblemareessential.Theobjectiveofthecallcontrol
problem is, given a spectrum of available frequencies
and users that wish tocommunicate, to maximize the benefit,
i.e., the number of users that communicate without
signalinterference.Weconsidercellularnetworksofreuse
distance k ≥ 2 and we study the online version of the
problem using competitive analysis. In cellular networks
of reuse distance 2, the previously best known algorithm
that beats the lower bound of 3 on the competitiveness
of deterministic algorithms, works on networks with one
frequency, achieves a competitive ratio against oblivious
adversaries, which is between 2.469 and 2.651, and uses
a number of random bits at least proportional to the size
of the network.We significantly improve this result by presentingaseriesofsimplerandomizedalgorithmsthathave
competitiveratiossignificantlysmallerthan3,workonnetworks
with arbitrarily many frequencies, and use only a
constant number of random bits or a comparable weak
random source. The best competitiveness upper bound
we obtain is 16/7 using only four random bits. In cellular
networks of reuse distance k > 2, we present simple
randomized online call control algorithms with competitive
ratios, which significantly beat the lower bounds on
the competitiveness of deterministic ones and use only
O(log k )randombits. Also,weshownewlowerboundson
thecompetitivenessofonlinecallcontrolalgorithmsincellularnetworksofanyreusedistance.
Inparticular,weshow
thatnoonline algorithm can achieve competitive ratio better
than 2, 25/12, and 2.5, in cellular networks with reuse
distancek ∈ {2, 3, 4},k = 5,andk ≥ 6, respectively.
Abstract: We address the issue of measuring distribution fairness in Internet-scale networks. This problem has several interesting instances encountered in different applications, ranging from assessing the distribution of load between network nodes for load balancing purposes, to measuring node utilization for optimal resource exploitation, and to guiding autonomous decisions of nodes in networks built with market-based economic principles. Although some metrics have been proposed, particularly for assessing load balancing algorithms, they fall short. We first study the appropriateness of various known and previously proposed statistical metrics for measuring distribution fairness. We put forward a number of required characteristics for appropriate metrics. We propose and comparatively study the appropriateness of the Gini coefficient (G) for this task. Our study reveals as most appropriate the metrics of G, the fairness index (FI), and the coefficient of variation (CV) in this order. Second, we develop six distributed sampling algorithms to estimate metrics online efficiently, accurately, and scalably. One of these algorithms (2-PRWS) is based on two effective optimizations of a basic algorithm, and the other two (the sequential sampling algorithm, LBS-HL, and the clustered sampling one, EBSS) are novel, developed especially to estimate G. Third, we show how these metrics, and especially G, can be readily utilized online by higher-level algorithms, which can now know when to best intervene to correct unfair distributions (in particular, load imbalances). We conclude with a comprehensive experimentation which comparatively evaluates both the various proposed estimation algorithms and the three most appropriate metrics (G, CV, andFI). Specifically, the evaluation quantifies the efficiency (in terms of number of the messages and a latency indicator), precision, and accuracy achieved by the proposed algorithms when estimating the competing fairness metrics. The central conclusion is that the proposed metric, G, can be estimated with a small number of messages and latency, regardless of the skew of the underlying distribution.
Abstract: For a place that gathers millions of people theWeb seems
pretty lonely at times. This is mainly due to the current predominant
browsing scenario; that of an individual participating
in an autonomous surfing session. We believe that
people should be seen as an integral part of the browsing
and searching activity towards a concept known as social
navigation. In this work, we extend the typical web
browser¢s functionality so as to raise awareness of other
people having similar web surfing goals at the current moment.
We further present features and algorithms that facilitate
online communication and collaboration towards common
searching targets. The utility of our system is established
by experimental studies. The extensions we present
can be easily adopted in a typical web browser.
Abstract: Online and Realtime counting and estimating the cardinality of sets is highly desirable for a large variety of applications, representing a foundational block for the efficient deployment and access of emerging internet scale information systems. In this work we implement three well known duplicate
insensitive counting algorithms and evaluate their performance in a testbed of resource-limited commercial off-the-shelf hardware devices. We focus on devices that can be used in wireless mobile and sensor applications and evaluate the memory complexity, time complexity and absolute error of the algorithms under different realistic scenaria. Our findings indicate the suitability of each algorithm depending on the application characteristics.
Abstract: We consider the online impairment-aware routing
and wavelength assignment (IA-RWA) problem in transparent
WDM networks. To serve a new connection, the online algorithm,
in addition to finding a route and a free wavelength (a lightpath),
has to guarantee its transmission quality, which is affected by
physical-layer impairments. Due to interference effects, the establishment
of the new lightpath affects and is affected by the other
lightpaths. We present two multicost algorithms that account
for the actual current interference among lightpaths, as well as
for other physical effects, performing a cross-layer optimization
between the network and physical layers. In multicost routing,
a vector of cost parameters is assigned to each link, from which
the cost vectors of the paths are calculated. The first algorithm
utilizes cost vectors consisting of impairment-generating source
parameters, so as to be generic and applicable to different physical
settings. These parameters are combined into a scalar cost
that indirectly evaluates the quality of candidate lightpaths. The
second algorithm uses specific physical-layer models to define
noise variance-related cost parameters, so as to directly calculate
the -factor of candidate lightpaths. The algorithms find a set of
so-called nondominated paths to serve the connection in the sense
that no path is better in the set with respect to all cost parameters.
To select the lightpath, we propose various optimization functions
that correspond to different IA-RWA algorithms. The proposed
algorithms combine the strength of multicost optimization with
low execution times, making them appropriate for serving online
connections
Abstract: In translucent (or managed reach) WDM optical
networks, regenerators are employed at specific nodes. Some of
the connections in such networks are routed transparently, while
others have to go through a sequence of 3R regenerators that serve
as “refueling stations” to restore their quality of transmission
(QoT). We extend an online multicost algorithm for transparent
networks presented in our previous study [1], to obtain an IA-RWA
algorithm that works in translucent networks and makes use,
when required, of the regenerators present at certain locations
of the network. To characterize a path, the algorithm uses a
multicost formulation with several cost parameters, including the
set of available wavelengths, the length of the path, the number of
regenerators used, and noise variance parameters that account for
the physical layer impairments. Given a new connection request
and the current utilization state of the network, the algorithm calculates
a set of non dominated candidate paths, meaning that any
path in this set is not inferior with respect to all cost parameters
than any other path. This set consists of all the cost-effective (in
terms of the domination relation) and feasible (in terms of QoT)
lightpaths for the given source-destination pair, including all the
possible combinations for the utilization of available regenerators
of the network. An optimization function or policy is then applied
to this set in order to select the optimal lightpath. Different optimization
policies correspond to different IA-RWA algorithms.
We propose and evaluate several optimization policies, such as the
most used wavelength, the best quality of transmission, the least
regeneration usage, or a combination of these rules. Our results
indicate that in a translucent network the employed IA-RWA
algorithm has to consider all problem parameters, namely, the
QoT of the lightpaths, the utilization of wavelengths and the
availability of regenerators, to efficiently serve the online traffic.
Abstract: We address the issue of measuring storage, or query load distribution fairness in peer-to-peer data management systems. Existing metrics may look promising from the point of view of specific peers, while in reality being far from optimal from a global perspective. Thus, first we define the requirements and study the appropriateness of various statistical metrics for measuring load distribution fairness towards these requirements. The metric proposed as most appropriate is the Gini coefficient (G). Second, we develop novel distributed sampling algorithms to compute G on-line, with high precision, efficiently, and scalably. Third, we show how G can readily be utilized on-line by higher-level algorithms which can now know when to best intervene to correct load imbalances. Our analysis and experiments testify for the efficiency and accuracy of these algorithms, permitting the online use of a rich and reliable metric, conveying a global perspective of the distribution.
Abstract: We propose and evaluate an impairment-aware multi-parametric routing and wavelength assignment algorithm for online traffic in transparent optical networks. In such networks the signal quality of transmission degrades due to physical layer impairments. In the multiparametric approach, a vector of cost parameters is assigned to each link, from which the cost vectors of candidate lightpaths are calculated. In the proposed scheme the cost vector includes impairment generating source parameters, such as the path length, the number of hops, the number of crosstalk sources and other inter-lightpath interfering parameters, so as to indirectly account for the physical layer effects. For a requested connection the algorithm calculates a set of candidate lightpaths, whose quality of transmission is validated using a function that combines the impairment generating parameters. For selecting the lightpath we propose and evaluate various optimization functions that correspond to different IA-RWA algorithms. Our performance results indicate that the proposed algorithms utilize efficiently the available resources and minimize the total accumulated signal degradation on the selected lightpaths, while having low execution times.
Abstract: This paper studies the data gathering problem in wireless networks, where data generated at the nodes has to be collected at a single sink. We investigate the relationship between routing optimality and fair resource management. In particular, we prove that for energy balanced data propagation, Pareto optimal routing and flow maximization are equivalent, and also prove that flow maximization is equivalent to maximizing the network lifetime. We algebraically characterize the network structures in which energy balanced data flows are maximal. Moreover, we algebraically characterize communication links which are not used by an optimal flow. This leads to the characterization of minimal network structures supporting the maximal flows.
We note that energy balance, although implying global optimality, is a local property that can be computed efficiently and in a distributed manner. We suggest online distributed algorithms for energy balance in different optimal network structures and numerically show their stability in particular setting. We remark that although the results obtained in this paper have a direct consequence in energy saving for wireless networks they do not limit themselves to this type of networks neither to energy as a resource. As a matter of fact, the results are much more general and can be used for any type of network and different type of resources.
Abstract: This paper studies the data gathering problem in wireless networks, where data generated at the nodes has to be collected at a single sink. We investigate the relationship between routing optimality and fair resource management. In particular, we prove that for energy-balanced data propagation, Pareto optimal routing and flow maximization are equivalent, and also prove that flow maximization is equivalent to maximizing the network lifetime. We algebraically characterize the network structures in which energy-balanced data flows are maximal. Moreover, we algebraically characterize communication links which are not used by an optimal flow. This leads to the characterization of minimal network structures supporting the maximal flows.
We note that energy-balance, although implying global optimality, is a local property that can be computed efficiently and in a distributed manner. We suggest online distributed algorithms for energy-balance in different optimal network structures and numerically show their stability in particular setting. We remark that although the results obtained in this paper have a direct consequence in energy saving for wireless networks they do not limit themselves to this type of networks neither to energy as a resource. As a matter of fact, the results are much more general and can be used for any type of network and different types of resources.
Abstract: We study a problem of scheduling client requests to servers. Each client has a particular latency requirement at each server and may choose either to be assigned to some server in order to get serviced provided that her latency requirement is met, or not to participate in the assignment at all. From a global perspective, in order to optimize the performance of such a system, one would aim to maximize the number of clients that participate in the assignment. However, clients may behave selfishly in the sense that, each of them simply aims to participate in an assignment and get serviced by some server where her latency requirement is met with no regard to overall system performance. We model this selfish behavior as a strategic game, show how to compute pure Nash equilibria efficiently, and assess the impact of selfishness on system performance. We also show that the problem of optimizing performance is computationally hard to solve, even in a coordinated way, and present efficient approximation and onlinealgorithms.
Abstract: We study a problem of scheduling client requests to servers.
Each client has a particular latency requirement at each server and may
choose either to be assigned to some server in order to get serviced provided
that her latency requirement is met or not to participate in the
assignment at all. From a global perspective, in order to optimize the
performance of such a system, one would aim to maximize the number
of clients that participate in the assignment. However, clients may behave
selfishly in the sense that each of them simply aims to participate
in an assignment and get serviced by some server where her latency requirement
is met with no regard to the overall system performance. We
model this selfish behavior as a strategic game, show how to compute
equilibria efficiently, and assess the impact of selfishness on system performance.
We also show that the problem of optimizing performance is
computationally hard to solve, even in a coordinated way, and present
efficient approximation and onlinealgorithms.
Abstract: We present an improved upper bound on the competitiveness of the online colouring algorithm First-Fit in disk graphs, which are graphs representing overlaps of disks on the plane. We also show that this bound is best possible for deterministic online colouring algorithms that do not use the disk representation of the input graph. We also present a related new lower bound for unit disk graphs.
Abstract: In this paper we describe a new simulation platform for complex wireless sensor networks that operate a collection of distributed algorithms and network protocols. Simulating such systems is complicated because of the need to coordinate different network layers and debug protocol stacks, often with very different interfaces, options, and fidelities. Our platform (which we call WSNGE) is a flexible and extensible environment that provides a highly scalable simulator with unique characteristics. It focuses on user friendliness, providing every function in both scriptable and visual way, allowing the researcher to define simulations and view results in an easy to use graphical environment. Unlike other solutions, WSNGE does not distinguish between different scenario types, allowing multiple different protocols to run at the same time. It enables rich online interaction with running simulations, allowing parameters, topologies or the whole scenario to be altered at any point in time.