Showing papers on "Load balancing (computing) published in 2005"

Live migration of virtual machines

Abstract: Migrating operating system instances across distinct physical hosts is a useful tool for administrators of data centers and clusters: It allows a clean separation between hard-ware and software, and facilitates fault management, load balancing, and low-level system maintenance.By carrying out the majority of migration while OSes continue to run, we achieve impressive performance with minimal service downtimes; we demonstrate the migration of entire OS instances on a commodity cluster, recording service downtimes as low as 60ms. We show that that our performance is sufficient to make live migration a practical tool even for servers running interactive loads.In this paper we consider the design options for migrating OSes running services with liveness constraints, focusing on data center and cluster environments. We introduce and analyze the concept of writable working set, and present the design, implementation and evaluation of high-performance OS migration built on top of the Xen VMM.

A survey and comparison of peer-to-peer overlay network schemes

Abstract: Over the Internet today, computing and communications environments are significantly more complex and chaotic than classical distributed systems, lacking any centralized organization or hierarchical control. There has been much interest in emerging Peer-to-Peer (P2P) network overlays because they provide a good substrate for creating large-scale data sharing, content distribution, and application-level multicast applications. These P2P overlay networks attempt to provide a long list of features, such as: selection of nearby peers, redundant storage, efficient search/location of data items, data permanence or guarantees, hierarchical naming, trust and authentication, and anonymity. P2P networks potentially offer an efficient routing architecture that is self-organizing, massively scalable, and robust in the wide-area, combining fault tolerance, load balancing, and explicit notion of locality. In this article we present a survey and comparison of various Structured and Unstructured P2P overlay networks. We categorize the various schemes into these two groups in the design spectrum, and discuss the application-level network performance of each group.

Multiple controlled mobile elements (data mules) for data collection in sensor networks

Abstract: Recent research has shown that using a mobile element to collect and carry data mechanically from a sensor network has many advantages over static multihop routing. We have an implementation as well employing a single mobile element. But the network scalability and traffic may make a single mobile element insufficient. In this paper we investigate the use of multiple mobile elements. In particular, we present load balancing algorithm which tries to balance the number of sensor nodes each mobile element services. We show by simulation the benefits of load balancing.

Cross-layer design for lifetime maximization in interference-limited wireless sensor networks

Abstract: We consider the joint optimal design of physical, medium access control (MAC), and routing layers to maximize the lifetime of energy-constrained wireless sensor networks. The problem of computing a lifetime-optimal routing flow, link schedule, and link transmission powers is formulated as a non-linear optimization problem. We first restrict the link schedules to the class of interference-free time division multiple access (TDMA) schedules. In this special case we formulate the optimization problem as a mixed integer-convex program, which can be solved using standard techniques. For general non-orthogonal link schedules, we propose an iterative algorithm that alternates between adaptive link scheduling and computation of optimal link rates and transmission powers for a fixed link schedule. The performance of this algorithm is compared to other design approaches for several network topologies. The results illustrate the advantages of load balancing, multihop routing, frequency reuse, and interference mitigation in increasing the lifetime of energy-constrained networks. We also describe a partially distributed algorithm to compute optimal rates and transmission powers for a given link schedule.

Load balancing technique implemented in a data network device utilizing a data cache

Abstract: Techniques for implementing a load balanced server system are described which may be used for effecting electronic commerce over a data network. The system comprises a load balancing system and a plurality of servers in communication with the load balancing system. Each of the plurality of servers may include a respective data cache for storing state information relating to client session transactions conducted between the server and a particular client. The load balancing system can be configured to select, using a load balancing protocol, an available first server from the plurality of servers to process an initial packet received from a source device such as, for example, a client machine of a customer. The load balancing system can also configured to route subsequent packets received from the source device to the first server. Before generating its response, the first server may verify that the state information relating to a specific client session stored in the data cache is up-to-date. If the first server determines that the state information stored in the data cache is not up-to-date, then the first server may be configured to retrieve the desired up-to-date state information from a database which is configured to store all state information relating to client sessions which have been initiated with the server system.

Parallel intrusion detection sensors with load balancing for high speed networks

Abstract: Various embodiments of a method and system for detecting unauthorized signatures to or from a local network. Multiple sensors are connected at an internetworking device, which can be a router or a switch. The sensors operate in parallel and each receives a portion of traffic through the internetworking device, at a session-based level or at a lower (packet-based) level. Depending on the type of internetworking device (router or switch) the load balancing mechanism that distributes the packets can be internal or external to the internetworking device. Also depending on the level of packet distribution (session-based or packet-based), the sensors share a network analyzer (if session-based) or both a network analyzer and a session analyzer (if packet-based).

Load balancing using distributed forwarding agents with application based feedback for different virtual machines

Abstract: A system and method are described for selecting a server to handle a connection. The method includes receiving at a service manager a connection request intercepted by a network device having a forwarding agent that is operative to receive instructions from a service manager, the connection request having been forwarded from the forwarding agent on the network device to the service manager. A preferred server is selected at the service manager from among a group of available servers. The preferred server is the server that is to service the connection request. Instructions are sent from the service manager to the forwarding agent. The instructions include the preferred server that is to service the connection request so that the connection request may be forwarded from the network device to the preferred server.

Systems and methods for performing multi-path storage operations

Abstract: Systems and methods for allocating transmission resources within a computer network are provided. In some embodiments of the invention, communication links may be assigned based on predefined preferences or system configuration to facilitate the transfer of data from one point in the network to another. In other embodiments, system operation may be monitored and communication paths be assigned dynamically based on this information to improve system operation and provide improved failover response, load balancing and to promote robust data access via alternative routes.

Multiple instance spanning tree protocol

Abstract: A multiple instance spanning tree protocol (MI-STP) creates a plurality of active topologies (i.e., loop-free paths) within a computer network. These active topologies may be established through the exchange and processing of multiple instance spanning tree bridge protocol data unit messages (MI-STP BPDUs) by the intermediate network devices within the network. The active topologies are preferably created independently of any virtual local area network (VLAN) designations defined within the network. Once the active topologies are defined, each VLAN designation is then mapped to a single active topology, although multiple VLAN designations are preferably mapped to the same active topology to provide load balancing.

Interference-aware Load Balancing for Multihop Wireless Networks

Abstract: Load balancing is critical for improving performance in wireless mesh networks. The unique characteristics of mesh networks, such as static nodes and the shared nature of the wireless medium, invalidate existing solutions from both wired and wireless networks and introduce new challenges for providing load balancing. In this paper, we focus on addressing these challenges. We first formulate the objective of load balancing in mesh networks and provide a theoretical solution to optimally achieve this objective. Then, we investigate some existing practical approaches to load balancing in mesh networks and show that none of them sufficiently address these challenges and some may even cause non-optimal paths and forwarding loops. In response, we propose a new path weight function, called MIC, and a novel routing scheme, called LIBRA, to provide interferenceaware and multi-channel/multi-radio aware load balancing for mesh networks, while still ensuring routing optimality and loopfreedom. We use extensive simulations to evaluate our scheme by comparing it with both the theoretical optimal solution and existing practical solutions. The results show close-to-optimum performance and indicate that LIBRA is a good candidate for load balancing and routing in wireless mesh networks.

Performance management for cluster-based web services

Abstract: We present an architecture and prototype implementation of a performance management system for cluster-based web services. The system supports multiple classes of web services traffic and allocates server resources dynamically so to maximize the expected value of a given cluster utility function in the face of fluctuating loads. The cluster utility is a function of the performance delivered to the various classes, and this leads to differentiated service. In this paper, we will use the average response time as the performance metric. The management system is transparent: it requires no changes in the client code, the server code, or the network interface between them. The system performs three performance management tasks: resource allocation, load balancing, and server overload protection. We use two nested levels of management. The inner level centers on queuing and scheduling of request messages. The outer level is a feedback control loop that periodically adjusts the scheduling weights and server allocations of the inner level. The feedback controller is based on an approximate first-principles model of the system, with parameters derived from continuous monitoring. We focus on SOAP-based web services. We report experimental results that show the dynamic behavior of the system.

Data traffic load balancing based on application layer messages

Abstract: A method is disclosed for application layer message-based load balancing. According to one aspect, when a network element receives one or more data packets that collectively contain an application layer message, the network element determines a message classification to which the application layer message belongs. Using a load-balancing algorithm that is mapped to the message classification, the network element selects a server from among a plurality of servers, and sends the message toward that server. According to one 'adaptive' load-balancing algorithm, the network element selects the server based on multiple servers' average historical response times and average outstanding request wait times. The network element continuously maintains these statistics for each server toward which the network element has sent requests. The network element tracks response times by recording how much time passes between the sending of a request to a server and the receiving of a corresponding response from that server.

Efficient, proximity-aware load balancing for DHT-based P2P systems

Abstract: Many solutions have been proposed to tackle the load balancing issue in DHT-based P2P systems. However, all these solutions either ignore the heterogeneity nature of the system, or reassign loads among nodes without considering proximity relationships, or both. In this paper, we present an efficient, proximity-aware load balancing scheme by using the concept of virtual servers. To the best of our knowledge, this is the first work to use proximity information in load balancing. In particular, our main contributions are: 1) relying on a self-organized, fully distributed k-ary tree structure constructed on top of a DHT, load balance is achieved by aligning those two skews in load distribution and node capacity inherent in P2P systems - that is, have higher capacity nodes carry more loads; 2) proximity information is used to guide virtual server reassignments such that virtual servers are reassigned and transferred between physically close heavily loaded nodes and lightly loaded nodes, thereby minimizing the load movement cost and allowing load balancing to perform efficiently; and 3) our simulations show that our proximity-aware load balancing scheme reduces the load movement cost by 11-65 percent for all the combinations of two representative network topologies, two node capacity profiles, and two load distributions of virtual servers. Moreover, we achieve virtual server reassignments in O(log N) time.

Method and apparatus for routing data to a load balanced server using MPLS packet labels

Abstract: A method of routing data to a load-balanced server through a network having one or more load-balancing nodes is disclosed, comprising receiving a label value; storing the label value in a load balancing mapping at a load-balancing node in a network, wherein the load balancing mapping associates the label with a packet flow and with interface identifying information; and forwarding subsequent packets of the flow to a selected load-balancing server. The forwarding route is defined by the load-balancing mapping and without hop-by-hop routing decisions. The first server response packet is switched hop-by-hop and the label is stored at each node traversed by the response packets, with a flow identifier and interface identifiers. For other request and response packets, nodes fast-switch the packets based on the label mappings; thus, packet flows are rapidly routed from the client to the same server without time-consuming hop-by-hop routing or repeated load-balancing decisions.

SMART: a scan-based movement-assisted sensor deployment method in wireless sensor networks

Abstract: The efficiency of sensor networks depends on the coverage of the monitoring area. Although in general a sufficient number of sensors are used to ensure a certain degree of redundancy in coverage so that sensors can rotate between active and sleep modes, a good sensor deployment is still necessary to balance the workload of sensors. In a sensor network with locomotion facilities, sensors can move around to self-deploy. The movement-assisted sensor deployment deals with moving sensors from an initial unbalanced state to a balanced state. Therefore, various optimization problems can be defined to minimize different parameters, including total moving distance, total number of moves, communication/computation cost, and convergence rate. In this paper, we propose a Scan-based Movement-Assisted sensoR deploymenT method (SMART) that uses scan and dimension exchange to achieve a balanced state. SMART also addresses a unique problem called communication holes in sensor networks. Using the concept of load balancing, SMART achieves good performance especially when applied to uneven distribution sensor networks, and can be a complement to the existing sensor deployment methods. Extensive simulation has been done to verify the effectiveness of the proposed scheme.

Session load balancing and use of VIP as source address for inter-cluster traffic through the use of a session identifier

Abstract: A method and system for enhancing a load balancing network's ability to load balance sessions is presented. A session identifier is placed within the TCP packet to enable a new mechanism of load distribution and connection grouping within a load balancing system. Specifically, TCP is invoked by a user application to obtain a unique session identifier value. TCP places the session identifier as a sender session identifier within the OPTIONS field of the TCP header. As an alternative, a session identifier can be placed within a session header that is attached to the TCP packet. On receiving such a packet, the destination load balancing system hashes over at least the session identifier value, and the node corresponding to the results of the hash algorithm acquires the packet. This method of hashing ensures that the same node acquires all subsequently received TCP packets possessing the same session identifier regardless of the source IP address or source port information. The node then places an identical session identifier value in the form of a response session identifier in its TCP response packet. When the TCP packet with response session identifier is received at a load balancing system, the node whose session identifier matches the response session identifier acquires it. This ensures that a response to a node assigned to a virtual IP address is received by the proper node, thus preventing other member nodes assigned to the same VIP address from picking up the packet instead.

Heterogeneity and load balance in distributed hash tables

Abstract: Existing solutions to balance load in DHTs incur a high overhead either in terms of routing state or in terms of load movement generated by nodes arriving or departing the system. In this paper, we propose a set of general techniques and use them to develop a protocol based on Chord, called Y/sub 0/, that achieves load balancing with minimal overhead under the typical assumption that the load is uniformly distributed in the identifier space. In particular, we prove that Y/sub 0/ can achieve near-optimal load balancing, while moving little load to maintain the balance and increasing the size of the routing tables by at most a constant factor. Using extensive simulations based on real-world and synthetic capacity distributions, we show that Y/sub 0/ reduces the load imbalance of Chord from O(log n) to a less than 3.6 without increasing the number of links that a node needs to maintain. In addition, we study the effect of heterogeneity on both DHTs, demonstrating significantly reduced average route length as node capacities become increasingly heterogeneous. For a real-word distribution of node capacities, the route length in Y/sub 0/ is asymptotically less than half the route length in the case of a homogeneous system.

Distributed load balancing for single entry-point systems

Abstract: A method and system for distributing work load in a cluster of at least two service resources. Depending upon the configuration, a service resource may be an individual process, such as a single instance of a computer game, or a node on which multiple processes are executing, such as a Server. Initial connection requests from new clients are directed to a single entry-point service resource in the cluster, called an intake. A separate intake is designated for each type of service provided by the cluster. The clients are processed in a group at the service resource currently designated as the intake to which clients initially connected, for the duration of the session. Based upon its loading, the current intake service resource determines that another service resource in the cluster should become a new intake for subsequent connection requests received from new clients. Selection of another service resource to become the new intake is based on the current work load of each resource in the cluster. All resources in the cluster are periodically informed of the resource for each service being provided that was last designated as the intake, and of the current load on each resource in the cluster. Subsequently, new clients requesting a service are directed to the newly designated intake for that service and processed on that resource for the duration of the session by those clients.

Workload-aware load balancing for clustered Web servers

Abstract: We focus on load balancing policies for homogeneous clustered Web servers that tune their parameters on-the-fly to adapt to changes in the arrival rates and service times of incoming requests. The proposed scheduling policy, ADAPTLOAD, monitors the incoming workload and self-adjusts its balancing parameters according to changes in the operational environment such as rapid fluctuations in the arrival rates or document popularity. Using actual traces from the 1998 World Cup Web site, we conduct a detailed characterization of the workload demands and demonstrate how online workload monitoring can play a significant part in meeting the performance challenges of robust policy design. We show that the proposed load, balancing policy based on statistical information derived from recent workload history provides similar performance benefits as locality-aware allocation schemes, without requiring locality data. Extensive experimentation indicates that ADAPTLOAD results in an effective scheme, even when servers must support both static and dynamic Web pages.

System and method for real-time balancing of user workload across multiple storage systems with shared back end storage

Abstract: A system and method for real-time load balancing of user workload across a plurality of physical storage systems with shared back-end storage is provided. A load balancing process tracks usage metrics and determines a source and destination physical storage system and a virtual storage system (vfiler) to be migrated.

Locality aware dynamic load management for massively multiplayer games

Abstract: Most massively multiplayer game servers employ static partitioning of their game world into distinct mini-worlds that are hosted on separate servers. This limits cross-server interactions between players, and exposes the division of the world to players. We have designed and implemented an architecture in which the partitioning of game regions across servers is transparent to players and interactions are not limited to objects in a single region or server. This allows a finer grain partitioning, which combined with a dynamic load management algorithm enables us to better handle transient crowding by adaptively dispersing or aggregating regions from servers in response to quality of service violations.Our load balancing algorithm is aware of the spatial locality in the virtual game world. Based on localized information, the algorithm balances the load and reduces the cross server communication, while avoiding frequent reassignment of regions. Our results show that locality aware load balancing reduces the average user response time by up to a factor of 6 compared to a global algorithm that does not consider spatial locality and by up to a factor of 8 compared to static partitioning.

Load balancing for parallel forwarding

Abstract: Workload distribution is critical to the performance of network processor based parallel forwarding systems. Scheduling schemes that operate at the packet level, e.g., round-robin, cannot preserve packet-ordering within individual TCP connections. Moreover, these schemes create duplicate information in processor caches and therefore are inefficient in resource utilization. Hashing operates at the flow level and is naturally able to maintain per-connection packet ordering; besides, it does not pollute caches. A pure hash-based system, however, cannot balance processor load in the face of highly skewed flow-size distributions in the Internet; usually, adaptive methods are needed. In this paper, based on measurements of Internet traffic, we examine the sources of load imbalance in hash-based scheduling schemes. We prove that under certain Zipf-like flow-size distributions, hashing alone is not able to balance workload. We introduce a new metric to quantify the effects of adaptive load balancing on overall forwarding performance. To achieve both load balancing and efficient system resource utilization, we propose a scheduling scheme that classifies Internet flows into two categories: the aggressive and the normal, and applies different scheduling policies to the two classes of flows. Compared with most state-of-the-art parallel forwarding schemes, our work exploits flow-level Internet traffic characteristics.

Analysis of Load Balancing Techniques in Cloud Computing

Abstract: Cloud Computing is an emerging computing paradigm. It aims to share data, calculations, and service transparently over a scalable network of nodes. Since Cloud computing stores the data and disseminated resources in the open environment. So, the amount of data storage increases quickly. In the cloud storage, load balancing is a key issue. It would consume a lot of cost to maintain load information, since the system is too huge to timely disperse load. Load balancing is one of the main challenges in cloud computing which is required to distribute the dynamic workload across multiple nodes to ensure that no single node is overwhelmed. It helps in optimal utilization of resources and hence in enhancing the performance of the system. A few existing scheduling algorithms can maintain load balancing and provide better strategies through efficient job scheduling and resource allocation techniques as well. In order to gain maximum profits with optimized load balancing algorithms, it is necessary to utilize resources efficiently. This paper discusses some of the existing load balancing algorithms in cloud computing and also their challenges.

Cooperative multicast for maximum network lifetime

Abstract: We consider cooperative data multicast in a wireless network with the objective to maximize the network lifetime. We present the maximum lifetime accumulative broadcast (MLAB) algorithm that specifies the nodes' order of transmission and transmit power levels. We prove that the solution found by MLAB is optimal but not necessarily unique. The power levels found by the algorithm ensure that the lifetimes of the active relays are the same, causing them to fail simultaneously. For the same battery levels at all the nodes, the optimum transmit powers become the same. The simplicity of the solution is made possible by allowing the nodes that are out of the transmission range of a transmitter to collect the energy of unreliably received overheard signals. As a message is forwarded through the network, nodes will have multiple opportunities to reliably receive the message by collecting energy during each retransmission. We refer to this cooperative strategy as accumulative multicast. Cooperative multicast not only increases the multicast energy-efficiency by allowing for more energy radiated in the network to be collected, but also facilitates load balancing by relaxing the constraint that a relay has to transmit with power sufficient to reach its most disadvantaged child. When the message is to be delivered to all network nodes this cooperative strategy becomes accumulative broadcast (Maric and Yates, 2002). Simulation results demonstrate that cooperative broadcast significantly increased network lifetime compared with conventional broadcast. We also present the distributed MLAB algorithm for accumulative broadcast that determines the transmit power levels locally at the nodes.

Load balancing based on cache content

Abstract: A method, schema, and computer-readable medium provide various means for load balancing computing devices in a multi-server environment. The method, schema, and computer-readable medium for load balancing computing devices in a multi-server environment may be utilized in a networked server environment, implementing a spreadsheet application for manipulating a workbook, for example. The method, schema, and computer-readable medium operate to load balance computing devices in a multi-server environment including determining whether a file, such as a spreadsheet application workbook, resides in the cache of a particular server, such as a calculation server. Upon meeting certain conditions, the user request may be directed to the particular server.