Results 1 - 10
of
21
Multi-resource allocation: Fairness-efficiency tradeoffs in a unifying framework
- in Proc. IEEE INFOCOM
, 2012
"... Abstract—Quantifying the notion of fairness is under-explored when there are multiple types of resources and users request different ratios of the different resources. A typical example is datacenters processing jobs with heterogeneous resource requirements on CPU, memory, network, bandwidth, etc. T ..."
Abstract
-
Cited by 37 (1 self)
- Add to MetaCart
Abstract—Quantifying the notion of fairness is under-explored when there are multiple types of resources and users request different ratios of the different resources. A typical example is datacenters processing jobs with heterogeneous resource requirements on CPU, memory, network, bandwidth, etc. This paper develops a unifying framework addressing the fairnessefficiency tradeoff in light of multiple types of resources. We develop two families of fairness functions that provide different tradeoffs, characterize the effect of user requests ’ heterogeneity, and prove conditions under which these fairness measures satisfy the Pareto efficiency, sharing incentive, and envy-free properties. Intuitions behind the analysis are explained in two visualizations of multi-resource allocation. We also investigate people’s fairness perceptions through an online survey of allocation preferences and provide a brief overview of related work on fairness.
An axiomatic theory of fairness in network resource allocation. Working paper
, 2009
"... Abstract — We present a set of five axioms for fairness measures in resource allocation. A family of fairness measures satisfying the axioms is constructed. Well-known notions such as α-fairness, Jain’s index, and entropy are shown to be special cases. Properties of fairness measures satisfying the ..."
Abstract
-
Cited by 25 (1 self)
- Add to MetaCart
Abstract — We present a set of five axioms for fairness measures in resource allocation. A family of fairness measures satisfying the axioms is constructed. Well-known notions such as α-fairness, Jain’s index, and entropy are shown to be special cases. Properties of fairness measures satisfying the axioms are proven, including Schur-concavity. Among the engineering implications is a generalized Jain’s index that tunes the resolution of the fairness measure, a new understanding of α-fair utility functions, and an interpretation of “larger α is more fair”. We also construct an alternative set of four axioms to capture efficiency objectives and feasibility constraints. I. QUANTIFYING FAIRNESS Given a vector x ∈ R n +, where xi is the resource allocated to user i, how fair is it? One approach to quantify the degree of fairness associated
A Foundation for Stochastic Bandwidth Estimation of Networks with Random Service
, 2010
"... We develop a stochastic foundation for bandwidth estimation of networks with random service, where bandwidth availability is expressed in terms of bounding functions with a defined violation probability. Exploiting properties of a stochastic max-plus algebra and system theory, the task of bandwidth ..."
Abstract
-
Cited by 6 (2 self)
- Add to MetaCart
We develop a stochastic foundation for bandwidth estimation of networks with random service, where bandwidth availability is expressed in terms of bounding functions with a defined violation probability. Exploiting properties of a stochastic max-plus algebra and system theory, the task of bandwidth estimation is formulated as inferring an unknown bounding function from measurements of probing traffic. We derive an estimation methodology that is based on iterative constant rate probes. Our solution provides evidence for the utility of packet trains for bandwidth estimation in the presence of variable cross traffic. Taking advantage of statistical methods, we show how our estimation method can be realized in practice, with adaptive train lengths of probe packets, probing rates, and replicated measurements required to achieve both high accuracy and confidence levels. We evaluate our method in a controlled testbed network, where we show the impact of cross traffic variability on the time-scales of service availability, and provide a comparison with existing bandwidth estimation tools.
Dominant Resource Fairness in Cloud Computing Systems with Heterogeneous Servers
, 2014
"... Cloud computing system represents unprecedented heterogeneity Server specification ..."
Abstract
-
Cited by 6 (2 self)
- Add to MetaCart
(Show Context)
Cloud computing system represents unprecedented heterogeneity Server specification
An Axiomatic Theory of Fairness in Resource Allocation
"... We present a set of five axioms for fairness measures in resource allocation: the axiom of continuity, of homogeneity, of saturation, of partition, and of starvation. We prove that there is a unique family of fairness measures satisfying the axioms, which is constructed and shown to include α-fairn ..."
Abstract
-
Cited by 5 (2 self)
- Add to MetaCart
(Show Context)
We present a set of five axioms for fairness measures in resource allocation: the axiom of continuity, of homogeneity, of saturation, of partition, and of starvation. We prove that there is a unique family of fairness measures satisfying the axioms, which is constructed and shown to include α-fairness, Jain’s index, and entropy as special cases. We prove properties of fairness measures satisfying the axioms, including symmetry and Schur-concavity. Among the engineering implications is a generalized Jain’s index that tunes the resolution of fairness measure, a decomposition of α-fair utility functions into fairness and efficiency components, and an interpretation of “larger α is more fair”. We further extend the axiomatic theory in three directions. First, the results are extended to quantify fairness of continuousdimension inputs, where resource allocations vary over time or domain. Second, by starting with both a vector of resource allocation and a vector of user-specific weights, and modifying the axiom of partition, we derive a new family of fairness measures that are asymmetric among users. Finally, a set of four axioms is developed by removing the axiom of homogeneity to capture a fairness-efficiency tradeoff. We present illustrative examples in congestion control, routing, power control, and spectrum management problems in communication networks, with the potential of a fairness evaluation tool explored. We also compare with other work of axiomatization in information, computer science, economics, sociology, and political philosophy.
Basic Theorems on the Backoff Process in 802.11 ∗
"... Since its introduction, the performance of IEEE 802.11 has attracted a lot of research attention and the center of the attention has been the throughput. For throughput analysis, in the seminal paper by Kumar et al. [8], they axiomized several ..."
Abstract
-
Cited by 3 (0 self)
- Add to MetaCart
(Show Context)
Since its introduction, the performance of IEEE 802.11 has attracted a lot of research attention and the center of the attention has been the throughput. For throughput analysis, in the seminal paper by Kumar et al. [8], they axiomized several
AP association for proportional fairness in multi-rate WLANs
- IEEE/ACM Transactions on Networking
"... Abstract—In this study, we investigate the problem of achieving proportional fairness via AP association in multi-rate WLANs. This problem is formulated as a non-linear programming with an objective function of maximizing the total user bandwidth utilities in the whole network. Such a formulation jo ..."
Abstract
-
Cited by 3 (1 self)
- Add to MetaCart
(Show Context)
Abstract—In this study, we investigate the problem of achieving proportional fairness via AP association in multi-rate WLANs. This problem is formulated as a non-linear programming with an objective function of maximizing the total user bandwidth utilities in the whole network. Such a formulation jointly considers fair-ness and AP selection. We first propose a centralized algorithm NLAO-PF to derive the user-AP association via relaxation. Since the relaxation may cause a large integrality gap, a compensation function is introduced to ensure that our algorithm can achieve at least half of the optimal in the worst-case. This algorithm is assumed to be adopted periodically for resource management. To handle the case of dynamic user membership, we propose a distributed heuristic BPF based on a novel performance revenue function, which provides an AP selection criterion for newcomers. When an existing user leaves the network, the transmission times of other users associated with the same AP can be redistributed easily based on NLAO-PF. Extensive simulation study has been performed to validate our design and to compare the performance of our algorithms with those of the state-of-the-art. Index Terms—AP association; bandwidth allocation; multi-rate WLANs; proportional fairness.
A Guide to the Stochastic Network Calculus
"... Abstract—The aim of the stochastic network calculus is to comprehend statistical multiplexing and scheduling of non-trivial traffic sources in a framework for end-to-end analysis of multi-node networks. To date, several models, some of them with subtle yet important differences, have been explored t ..."
Abstract
-
Cited by 3 (2 self)
- Add to MetaCart
Abstract—The aim of the stochastic network calculus is to comprehend statistical multiplexing and scheduling of non-trivial traffic sources in a framework for end-to-end analysis of multi-node networks. To date, several models, some of them with subtle yet important differences, have been explored to achieve these objectives. Capitalizing on previous works, this paper contributes an intuitive approach to the stochastic network calculus, where we seek to obtain its fundamental results in the possibly easiest way. For this purpose, we will now and then trade generality or precision for simplicity. In detail, the method that is assembled in this work uses moment generating functions, known from the theory of effective bandwidths, to characterize traffic arrivals and network service. Thereof, affine envelope functions with exponentially decaying overflow profile are derived to compute statistical end-to-end backlog and delay bounds for networks. I.
Approximate optimization for proportional fair AP association in multi-rate WLANs
- in 5th Int. Cof. WASA
, 2010
"... Abstract. In this study, we investigate the problem of achieving proportional fairness via Access Point (AP) association in multi-rate WLANs. This problem is formulated as a non-linear program with an objective function of maximizing the total user bandwidth utilities in the whole network. It is NP- ..."
Abstract
-
Cited by 3 (2 self)
- Add to MetaCart
(Show Context)
Abstract. In this study, we investigate the problem of achieving proportional fairness via Access Point (AP) association in multi-rate WLANs. This problem is formulated as a non-linear program with an objective function of maximizing the total user bandwidth utilities in the whole network. It is NP-hard, and therefore effort in this paper is made to seek approximate solutions. We propose a centralized algorithm to derive the user-AP association via relaxation. Such a relaxation may cause a large integrality gap. Therefore a compensation function is introduced to guarantee that our algorithm can achieve at least half of the optimal solution in the worst-case scenario theoretically. Extensive simulation study has been reported to validate and compare the performances of our algorithms with those of the state-of-the-art.
