Cluster scheduling and load balancing via TCP options

This paper describes an experiment in designing, implementing and testing a Transport layer cluster scheduling and dispatching architecture. The motivation for the experiment was the hypothesis that a Transport layer clustering solution may offer advantantages over the existing industry-standard Network layer and Data Link Layer approaches. The critical success factors initially established to guide and evaluate the experiment were reduced dispatcher work load, reduced dispatcher internal state memory requirements, distributed denial of service resilience, and cluster software design simplicity. The functional design stage of the experiment produced a Transport layer strategy for scheduling and load balancing based on the specification of two new TCP options. Implementation required the introduction of the newly specified TCP options into the Linux (2.4) kernel. The implementation produced an extended Linux Socket API to facilitate user-process access to the additional TCP capability. The testing stage of the experiment confirmed the operational efficiency of the solution.

ELISA: An Estimated Load Information Scheduling Algorithm for Distributed Computing Systems

In this paper, we present a decentralized dynamic load scheduling/balancing algorithm called ELISA (Estimated Load Information Scheduling Algorithm) for general purpose distributed computing systems. ELISA uses estimated state information based upon periodic exchange of exact state information between neighbouring nodes to perform load scheduling. The primary objective of the algorithm is to cut down on the communication and load transfer overheads by minimizing the frequency of status exchange and by restricting the load transfer and status exchange within the buddy set of a processor. It is shown that the resulting algorithm performs almost as well as a perfect information algorithm and is superior to other load balancing schemes based on the random sharing and Ni-Hwang algorithms. A sensitivity analysis to study the effect of various design parameters on the effectiveness of load balancing is also carried out. Finally, the algorithm's performance is tested on large dimensional hypercubes in the presence of time-varying load arrival process and is shown to perform well in comparison to other algorithms. This makes ELISA a viable and implementable load balancing algorithm for use in general purpose distributed computing systems.

Adaptive optimal load balancing in a nonhomogeneous multiserver system with a central job scheduler

A model comprising several servers, each equipped with its own queue and with possibly different service speeds, is considered. Each server receives a dedicated arrival stream of jobs; there is also a stream of generic jobs that arrive to a job scheduler and can be individually allocated to any of the servers. It is shown that if the arrival streams are all Poisson and all jobs have the same exponentially distributed service requirements, the probabilistic splitting of the generic stream that minimizes the average job response time is such that it balances the server idle times in a weighted least-squares sense, where the weighting coefficients are related to the service speeds of the servers. The corresponding result holds for nonexponentially distributed service times if the service speeds are all equal. This result is used to develop adaptive quasi-static algorithms for allocating jobs in the generic arrival stream when the load parameters are unknown. The algorithms utilize server idle-time measurements which are sent periodically to the central job scheduler. A model is developed for these measurements, and the result mentioned is used to cast the problem into one of finding a projection of the root of an affine function, when only noisy values of the function can be observed

Load balancing algorithms for an extended hypercube

Reduction of the execution time of a job through equitable distribution of work load among the processors in a distributed system is the goal of load balancing. Performance of static and dynamic load balancing algorithms for the extended hypercube, is discussed. Threshold algorithms are very well-known algorithms for dynamic load balancing in distributed systems. An extension of the threshold algorithm, called the multilevel threshold algorithm, has been proposed. The hierarchical interconnection network of the extended hypercube is suitable for implementing the proposed algorithm. The new algorithm has been implemented on a transputer-based system and the performance of the algorithm for an extended hypercube is compared with those for mesh and binary hypercube networks

Relay Load Balancing in Queued Cooperative Wireless Networks with Rateless Codes

Relay selection combined with buffering of packets of relays can substantially increase the throughput of a cooperative network that uses rateless codes. However, buffering also increases the end-to-end delays due to the additional queuing delays at the relay nodes. In this paper we propose a novel method that exploits a unique property of rateless codes that enables a receiver to decode a packet from non-contiguous and unordered portions of the received signal. In it, each relay, depending on its queue length, ignores its received coded bits with a given probability. We show that this substantially reduces the end-to-end delays while retaining almost all of the throughput gain achieved by buffering. In effect, the method increases the odds that the packet is first decoded by a relay with a smaller queue. Thus, the queuing load is balanced across the relays and traded off with transmission times. We derive explicit necessary and sufficient conditions for the stability of this system when the various channels undergo fading. Despite encountering analytically intractable G/GI/1 queues in our system, we also gain insights about the method by analyzing a similar system with a simpler model for the relay-to-destination transmission times.

Distributed load adaptive scheduling for throughput and delay guarantees in fading wireless LANs

We consider a problem of providing mean delay and average throughput guarantees in random access fading wireless channels using CSMA/CA algorithm. This problem becomes much more challenging when the scheduling is distributed as is the case in a typical local area wireless network. We model the CSMA network using a novel queueing network based approach. The optimal throughput per device and throughput optimal policy in an M device network is obtained. We provide a simple contention control algorithm that adapts the attempt probability based on the network load and obtain bounds for the packet transmission delay. The information we make use of is the number of devices in the network and the queue length (delayed) at each device. The proposed algorithms stay within the requirements of the IEEE 802.11 standard.

Load Balancing with Availability Checker and Load Reporters (LB-ACLRs) for improved Performance in Distributed Systems

Distributed system has quite a lot of servers to attain increased availability of service and for fault tolerance. Balancing the load among these servers is an important task to achieve better performance. There are various hardware and software based load balancing solutions available. However there is always an overhead on Servers and the Load Balancer while communicating with each other and sharing their availability and the current load status information. Load balancer is always busy in listening to clients' request and redirecting them. It also needs to collect the servers' availability status frequently, to keep itself up-to-date. Servers are busy in not only providing service to clients but also sharing their current load information with load balancing algorithms. In this paper we have proposed and discussed the concept and system model for software based load balancer along with Availability-Checker and Load Reporters (LB-ACLRs) which reduces the overhead on server and the load balancer. We have also described the architectural components with their roles and responsibilities. We have presented a detailed analysis to show how our proposed Availability Checker significantly increases the performance of the system.

Simple Load Balancing for Distributed Hash Tables

Distributed hash tables have recently become a useful building block for a variety of distributed applications. However, current schemes based upon consistent hashing require both considerable implementation complexity and substantial storage overhead to achieve desired load balancing goals. We argue in this paper that these goals can b e achieved more simply and more cost-effectively. First, we suggest the direct application of the "power of two choices" paradigm, whereby an item is stored at the less loaded of two (or more) random alternatives. We then consider how associating a small constant number of hash values with a key can naturally b e extended to support other load balancing methods, including load-stealing or load-shedding schemes, as well as providing natural fault-tolerance mechanisms.

Load Balancing a Cluster of Web Servers using Distributed Packet Rewriting

In this paper, we propose and evaluate an implementation of a prototype scalable web server. The prototype consists of a load-balanced cluster of hosts that collectively accept and service TCP connections. The host IP addresses are advertised using the Round Robin DNS technique, allowing any host to receive requests from any client. Once a client attempts to establish a TCP connection with one of the hosts, a decision is made as to whether or not the connection should be redirected to a different host---namely, the host with the lowest number of established connections. We use the low-overhead Distributed Packet Rewriting (DPR) technique to redirect TCP connections. In our prototype, each host keeps information about connections in hash tables and linked lists. Every time a packet arrives, it is examined to see if it has to be redirected or not. Load information is maintained using periodic broadcasts amongst the cluster hosts.

An automatable generic strategy for dynamic load balancing in parallel structured mesh CFD codes

Parallel computing is now widely used in numerical simulation, particularly for application codes based on finite difference and finite element methods. A popular and successful technique employed to parallelize such codes onto large distributed memory systems is to partition the mesh into sub-domains that are then allocated to processors. The code then executes in parallel, using the SPMD methodology, with message passing for inter-processor interactions. In order to improve the parallel efficiency of an imbalanced structured mesh CFD code, a new dynamic load balancing (DLB) strategy has been developed in which the processor partition range limits of just one of the partitioned dimensions uses non-coincidental limits, as opposed to coincidental limits. The ‘local’ partition limit change allows greater flexibility in obtaining a balanced load distribution, as the workload increase, or decrease, on a processor is no longer restricted by the ‘global’ (coincidental) limit change. The automatic implementation of this generic DLB strategy within an existing parallel code is presented in this chapter, along with some preliminary results.

Shape-optimized mesh partitioning and load balancing for parallel adaptive FEM

We present a dynamic distributed load balancing algorithm for parallel, adaptive Finite Element simulations in which we use preconditioned Conjugate Gradient solvers based on domain-decomposition. The load balancing is designed to maintain good partition aspect ratio and we show that cut size is not always the appropriate measure in load balancing. Furthermore, we attempt to answer the question why the aspect ratio of partitions plays an important role for certain solvers. We define and rate different kinds of aspect ratio and present a new center-based partitioning method of calculating the initial distribution which implicitly optimizes this measure. During the adaptive simulation, the load balancer calculates a balancing flow using different versions of the diffusion algorithm and a variant of breadth first search. Elements to be migrated are chosen according to a cost function aiming at the optimization of subdomain shapes. Experimental results for Bramble's preconditioner and comparisons to state-of-the-art load balancers show the benefits of the construction.

Dynamic load-balancing of finite element applications with the DRAMA library

The DRAMA library, developed within the European Commission funded (ESPRIT) project DRAMA, supports dynamic load-balancing for parallel (message-passing) mesh-based applications. The target applications are those with dynamic and solution-adaptive features. The focus within the DRAMA project was on finite element simulation codes for structural mechanics. An introduction to the DRAMA library will illustrate that the very general cost model and the interface designed specifically for application requirements provide simplified and effective access to a range of parallel partitioners. The main body of the paper will demonstrate the ability to provide dynamic load-balancing for parallel FEM problems that include: adaptive meshing, re-meshing, the need for multi-phase partitioning.

Load-balancing for mesh-based applications on heterogeneous cluster computers

This paper discusses load-balancing issues when using heterogeneous cluster computers. There is a growing trend towards the use of commodity microprocessor clusters. Although today's microprocessors have reached a theoretical peak performance in the range of one GFLOPS/s, heterogeneous clusters of commodity processors are amongst the most challenging parallel systems to programme efficiently. We will outline an approach for optimising the performance of parallel mesh-based applications for heterogeneous cluster computers and present case studies with the GeoFEM code. The focus is on application cost monitoring and load balancing using the DRAMA library.