Load balancing via random local search in closed and open systems

@article{Ganesh2010LoadBV,
  title={Load balancing via random local search in closed and open systems},
  author={Ayalvadi J. Ganesh and Sarah Lilienthal and D. Manjunath and Alexandre Prouti{\`e}re and Florian Simatos},
  journal={ArXiv},
  year={2010},
  volume={abs/1002.4314}
}
In this paper, we analyze the performance of random load resampling and migration strategies in parallel server systems. Clients initially attach to an arbitrary server, but may switch servers independently at random instants of time in an attempt to improve their service rate. This approach to load balancing contrasts with traditional approaches where clients make smart server selections upon arrival (e.g., Join-the-Shortest-Queue policy and variants thereof). Load resampling is particularly… 
View on ACM
arxiv.org
24 Citations
Highly Influential Citations
1
Background Citations
11
Methods Citations
5

Figures from this paper

24 Citations

Citation Type

Citation Type

Distributed load balancing in heterogenous systems
TLDR
It is proved that when the user population grows large, pure NEs get closer to a Proportionally Fair (PF) allocation of users to servers, and the gap between equilibriums and this ideal allocation depending on user population is characterized.
Distributed Proportional Fair Load Balancing in Heterogenous Systems
TLDR
In open systems, where users randomly enter the system and leave upon service completion, it is established that the RLS algorithm stabilizes the system whenever this it at all possible under centralized load balancing schemes, i.e., it is throughput-optimal.
FCD: Fast-concurrent-distributed load balancing under switching costs and imperfect observations
TLDR
It is shown that the system reaches an c-Nash equilibrium in expected time O, which implies that the convergence rate is robust with large scale system(large user population), and is not affected by imperfect measurements of the server load.
Random Walk Based Sampling for Load Balancing in Multi-Server Systems
  • Dengwang Tang, V. Subramanian
  • Computer Science
    Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems
  • 2019
TLDR
This paper analyzes a variant of the power-of-d choices algorithm, where d servers are sampled through d independent non-backtracking random walks on a k-regular graph, and the stationary distribution of the system converges to the stationary solution of the ODE.
Derandomized Load Balancing using Random Walks on Expander Graphs
TLDR
Under this scheme, the dynamics of the queuing system converges to the same deterministic ordinary differential equation (ODE) for the power-of-d choices scheme, and the stationary distribution of the system converging to the fixed point of the ODE.
Energy-aware capacity scaling in virtualized environments with performance guarantees
On the Asymptotic Insensitivity of the Supermarket Model in Processor Sharing Systems
TLDR
This paper demonstrates that for this class of hyperexponential distributions of order 2 global attraction of the unique fixed point can still be established using monotonicity by picking a suitable state space and partial order and can be leveraged to prove asymptotic insensitivity within thisclass of distributions for other load balancing systems.
Randomized Work Stealing versus Sharing in Large-scale Systems with Non-exponential Job Sizes
TLDR
The main insight is that work stealing benefits significantly from having more variable job sizes and work sharing may become inferior to work stealing for loads as small as $1/2 + \epsilon$ for any $\epssilon > 0$.
Heavy-Traffic Limits for a Many-Server Queueing Network with Switchover
We study a multiclass Markovian queueing network with switchover across a set of many-server stations. New arrivals to each station follow a nonstationary Poisson process. Each job waiting in queue…
Tight Load Balancing Via Randomized Local Search
TLDR
This analysis is based on the crucial observation that performing destructive moves (reversals of RLS moves) cannot decrease the balancing time, which allows us to simplify problem instances and to ignore “inconvenient moves” in the analysis.
...
1
2
3
...

References

SHOWING 1-10 OF 27 REFERENCES
The Power of Two Choices in Randomized Load Balancing
TLDR
This work uses a limiting, deterministic model representing the behavior as n/spl rarr//spl infin/ to approximate the behavior of finite systems and provides simulations that demonstrate that the method accurately predicts system behavior, even for relatively small systems.
Stability of load balancing algorithms in dynamic adversarial systems
TLDR
The proof techniques are very different from the ones used by Muthukrishnan and Rajaraman, and yield a simpler proof and tighter bounds on the difference in loads, and apply to the problem of routing packets in adversarial systems.
Convergence time to Nash equilibrium in load balancing
We study the number of steps required to reach a pure Nash equilibrium in a load balancing scenario where each job behaves selfishly and attempts to migrate to a machine which will minimize its cost.…
Bounds for the convergence rate of randomized local search in a multiplayer load-balancing game
TLDR
Randomized Local Search is shown to show that it can be realised by a simple distributed network of users that act selfishly, have no central control and only interact via the effect they have on the cost functions of resources.
Task assignment with unknown duration
TLDR
It is found that under heavy-tailed workloads, TAGS can outperform all task assignment policies known to us by several orders of magnitude with respect to both mean response time and mean slowdown, provided the system load is not too high.
Distributed selfish load balancing
TLDR
A natural protocol for the agents is discussed which can be implemented in a strongly distributed setting, uses no central control, and has good convergence properties and a lower bound of Ω (max{log log m, n}) for the convergence time is given.
Path Selection and Multipath Congestion Control
TLDR
This paper investigates the potential benefits of coordinated congestion control for multipath data transfers, and shows that with RTT bias (as in TCP Reno), uncoordinated control can lead to inefficient equilibria, and introduces the notion of a Nash equilibrium.
Capacity of Wireless Data Networks with Intra- and Inter-Cell Mobility
  • S. Borst, A. Proutière, N. Hegde
  • Computer Science
    Proceedings IEEE INFOCOM 2006. 25TH IEEE International Conference on Computer Communications
  • 2006
TLDR
This paper analyzes networks with several interacting base stations, and specifically examines the capacity impact of intraand inter-cell mobility, showing that mobility tends to increase the capacity, not only in case of globally optimal scheduling, but also when each of the base stations operates according to a fair sharing policy.
The natural work-stealing algorithm is stable
TLDR
It is shown that the system is stable for any constant generation rate /spl lambda/<1 and for a wide class of functions f, and a quantitative description of the functions f which lead to stable systems is given.
Nashification and the Coordination Ratio for a Selfish Routing Game
TLDR
The first result is an O(nm2) time algorithm for Nashification, which can be used in combination with any approximation algorithm for the routing problem to compute a Nash equilibrium of the same quality and yields a PTAS for the computation of a best Nash equilibrium.
...
1
2
3
...