Replica Servers Research Articles

A Recursively-Adjusting Co-allocation scheme with a Cyber-Transformer in Data Grids

A co-allocation architecture was developed in order to enable parallel downloads of datasets from multiple servers. Several co-allocation strategies have been coupled and used to exploit rate differences among various client–server links and to address dynamic rate fluctuations by dividing files into multiple blocks of equal sizes. However, a major obstacle, the idle time of faster servers having to wait for the slowest server to deliver the final block, makes it important to reduce differences in finish times among replica servers. In this paper, we propose Recursively-Adjusting Co-Allocation, a dynamic co-allocation scheme for improving data transfer performance in Data Grids. The experimental results show that our approach can reduce the idle time spent waiting for the slowest server and decrease data transfer completion times. We developed Cyber-Transformer, a new toolkit with a friendly GUI interface that makes it easy for inexperienced users to manage replicas and download files in Data Grid environments. We also provide an effective scheme for reducing the cost of reassembling data blocks.

A segmentation-based fine-grained peer sharing technique for delivering large media files in content distribution networks

Delivering large media files over the Internet is a challenging task because it has some unique features that are different from delivering conventional Web documents. In this paper, we propose a fine-grained peer sharing technique for dealing with the problem in the context of content distribution networks. The key difference of the technique from conventional peer-to-peer systems is that the unit of peer sharing is not a complete media file, but at a finer granularity. By doing so, we improve the flexibility of replica servers for handling client requests. We analyze the storage requirement at replica servers and design a scheduling algorithm to coordinate the delivery process from multiple replica servers to a client. Our simulations show that the fine-grained peer sharing approach can reduce the initial latency of clients and the rejection rate of the system significantly over a simple peer sharing method.

Session based access control in geographically replicated Internet services

Performance critical services over Internet often rely on geographically distributed architectures of replicated servers. Content Delivery Networks (CDN) are a typical example where service is based on a distributed architecture of replica servers to guarantee resource availability and proximity to final users. In such distributed systems, network links are not dedicated, and may be subject to external traffic. This brings up the need to develop access control policies that adapt to network load changing conditions. Further, Internet services are mainly session based, thus an access control support must take into account a proper differentiation of requests and perform session based decisions while considering the dynamic availability of resources due to external traffic.In this paper we introduce a distributed architecture with access control capabilities at session aware access points. We consider two types of services characterized by different patterns of resource consumption and priorities. We formulate a Markov Modulated Poisson Decision Process for access control that captures the heterogeneity of multimedia services and the variable availability of resources due to external traffic. The proposed model is optimized by means of stochastic analysis, showing the impact of external traffic on service quality. The structural properties of the optimal solutions are studied and considered as the basis for the formulation of heuristics. The performance of the proposed heuristics is studied by means of simulations, showing that in some typical scenario they perform close to the optimum.

An Adaptive Load Balancing Algorithm for Service Composition

Service Composition enables a new way to create new services by assembling independent service components. One of the important goals in service composition is load balancing across service replicas. In this paper, a novel adaptive distributed load capacity based (LCB) algorithm is presented to solve the load balancing problem in service composition. LCB algorithm uses service routing mechanism to search services and transfer the application data. By self-adaptively calculating the load capacity (LC) metric, LCB algorithm can choose a reasonable service path. LC metric reflects the workload of the service node, and is adaptively adjusted according to the variations of the server's load. Compared with the previous researches, LCB algorithm needn't the global topology information and load information. It is much suitable for dynamic service composition in distributed systems. Simulations show that LCB algorithm performs well in terms of load balancing across the service replicas.

부분적 분산형 수동적 중복 알고리즘

본 논문에서는 메시지 전달 분산시스템서 결정적 서버를 위한 부분적 분산형 수동적 중복 알고리즘을 제안한다. 이 알고리즘은 반드시 주 서버가 아니더라도 임의의 보조서버가 자신이 수신한 클라이언트 요구에 대해 서버로부터 그 요구의 전달일련번호를 얻은 후, 그 보조서버가 직접 해당 요구를 처리하고, 이에 대한 다른 중복 서버들과의 조정에 대한 책임자 역할을 수행할 수 있도록 한다. 이러한 바람직한 특성 때문에, 제안된 알고리즘이 기존의 부하균등 기법과 결합된다면, 주 서버에의 급격한 부하 발생을 효율적으로 피할 수 있다. 따라서, 이 알고리즘은 기존의 수동적 중복 알고리즘보다 결정적 중복 서버 시스템에 대한 높은 확장성을 제공할 수 있다. 본 논문에서 수행한 시뮬레이션에서 기존 알고리즘에 비해 제안한 알고리즘이 클라이언트 요구 당 평균응답시간을 <TEX>$16.5\%{\~}52.3\%$</TEX>까지 줄일 수 있다는 것을 보여준다. This paper presents a partially decentralized passive replication algorithm for deterministic servers in message-passing distributed systems. The algorithm allows any backup server, not necessarily the primary server, to take responsibility for processing its received client request and coordinating with the other replica servers after obtaining the delivery sequence number of the request from the primary. Thanks to thus desirable feature, the algorithm with conventional load balancing techniques can efficiently avoid extreme load conditions on the primary. Therefore, it can provide better scalability of deterministic and replicated sewer systems than traditional passive replication algorithms. Simulation results indicate that the proposed algorithm can reduce <TEX>$16.5\%{\~}52.3\%$</TEX> of the average response time of a client request compared with the traditional ones.

A replication framework for program‐to‐program interaction across unreliable networks and its implementation in a servlet container

AbstractWe propose a service replication framework for unreliable networks. The service exhibits the same consistency guarantees about the order of execution of operation requests as its non‐replicated implementation. Such guarantees are preserved in spite of server replica failure or network failure (either between server replicas or between a client and a server replica), and irrespective of when the failure occurs. Moreover, the service guarantees that in the case when a client sends an ‘update’ request multiple times, there is no risk that the request be executed multiple times. No hypotheses about the timing retransmission policy of clients are made, e.g. the very same request might even arrive at different server replicas simultaneously. All of these features make the proposed framework particularly suitable for interaction between remote programs, a scenario that is gaining increasing importance. We discuss a prototype implementation of our replication framework based on Tomcat, a very popular Java‐based Web server. The prototype comes into two flavors: replication of HTTP client session data and replication of a counter accessed as a Web service. Copyright © 2005 John Wiley &amp; Sons, Ltd.

A pull‐based service replication protocol in mobile ad hoc networks

AbstractThe mobility of nodes in ad hoc networks can lead to frequent network partitioning. This partitioning disconnects many nodes from a centralised server. Being aware of a future partitioning can help to provide a continuous service availability for all mobile nodes. In this paper, we propose a service replication protocol consisting of three distributed algorithms. This protocol (i) predicts the occurrence of network partitioning, which allows to create service replicas in the future separate partitions and (ii) constructs a dynamic service replicas deployment scheme. Using linear‐time temporal logic, we show that each detection of network partitioning can be predicted before its occurrence. Our simulation results show that the proposed protocol guarantees service availability to all mobile nodes without incurring high cost. Copyright © 2005 AEIT.

Specifying and using intrusion masking models to process distributed operations

It is important for critical applications to provide critical services without any integrity or availability degradation in the presence of intrusions. This requirement can be satisfied by intrusion masking techniques under some situations. Compared with intrusion tolerance techniques, where some i ntegrity or availability degradations are usually caused, intrusion masking techniques use substantial replications to avoid such degradations. Existing intrusion masking techniques, such as the state machine approach, can effectively mask intrusions when processing requests from a client using a server replica group, but they are fairly limited in processing a (multi-stage) distributed operation across multiple server replica groups. As more and more applications (e.g., supply chain management, distributed banking) need to process distributed operations in an intrusion-masking fashion, it is in urgent need to overcome the limitations of existing intrusion masking techniques. In this paper, we specify and compose two intrusion-masking models for inter-replica-group distributed computing. Using these two models, a variety of applications can mask (numerous kinds of) intrusions. Our intrusion masking models overcome the limitations of existing intrusion masking techniques. The survivability of our intrusion-masking models is quantitatively analyzed. A simple yet practical implementation method of our intrusion-masking models is proposed and applied to build two intrusion-masking two-phase-commit (2PC) protocols, and the corresponding efficiency is analyzed. The two intrusion-masking 2PC protocols and the analysis results show that the proposed intrusion-masking models have good utility, practicality, and survivability. Finally, the composition methodology developed in this paper can also be used to develop other intrusion-masking distributed computing models.

End-to-end latency of a fault-tolerant CORBA infrastructure

This paper presents an evaluation of the end-to-end latency of a fault-tolerant CORBA infrastructure that we have implemented. The fault-tolerant infrastructure replicates the server applications using active, passive and semi-active replication, and maintains strong replica consistency of the server replicas. By analyses and by measurements of the running fault-tolerant infrastructure, we characterize the end-to-end latency under fault-free conditions. The main determining factor of the run-time performance of the fault-tolerant infrastructure is the Totem group communication protocol, which contributes to the end-to-end latency primarily in two ways: the delay in sending messages and the processing cost of the rotating token. To reduce the delay in sending messages for passive and semi-active replication, the position of the primary server replica on the Totem ring, the token rotation time, the processing time at the client, and the processing time at the server must be considered. For active replication, the presence of duplicate messages adversely affects the performance. However, if an effective sending-side duplicate suppression mechanism is implemented, active replication is more advantageous than both passive and semi-active replication because of the automatic selection of the most favorable position of the server replica that sends the first non-duplicate reply.

Minimizing Delivery Cost in Scalable Streaming Content Distribution Systems

Recent scalable multicast streaming protocols for on-demand delivery of media content offer the promise of greatly reduced server and network bandwidth. However, a key unresolved issue is how to design scalable content distribution systems that place replica servers closer to various client populations and route client requests and response streams so as to minimize the total server and network delivery cost. This issue is significantly more complex than the design of distribution systems for traditional Web files or unicast on-demand streaming, for two reasons. First, closest server and shortest path routing does not minimize network bandwidth usage; instead, the optimal routing of client requests and server multicasts is complex and interdependent. Second, the server bandwidth usage increases with the number of replicas. Nevertheless, this paper shows that the complex replica placement and routing optimization problem, in its essential form, can be expressed fairly simply, and can be solved for example client populations and realistic network topologies. The solutions show that the optimal scalable system can differ significantly from the optimal system for conventional delivery. Furthermore, simple canonical networks are analyzed to develop insights into effective heuristics for near-optimal placement and routing. The proposed new heuristics can be used for designing large and heterogeneous systems that are of practical interest. For a number of example networks, the best heuristics produce systems with total delivery cost that is within 16% of optimality.

Three‐tier replication for FT‐CORBA infrastructures

AbstractEnforcing strong replica consistency among a set of replicas of a service deployed across an asynchronous distributed system in the presence of crash failures is a real practical challenge. If each replica runs the consistency protocol bundled with the actual service implementation, this target cannot be achieved, as replicas need to be located over a partially synchronous distributed system to solve the distributed agreement problems underlying strong replica consistency.A three‐tier architecture for software replication enables the separation of the replication logic, i.e. protocols and mechanisms necessary for managing software replication, from both clients and server replicas. The replication logic is embedded in a middle‐tier that confines the need of partial synchrony and thus frees replica deployment.In this paper we first introduce the basic concepts underlying three‐tier replication. Then we present the interoperable replication logic (IRL) architecture, a fault‐tolerant CORBA compliant infrastructure. IRL exploits a three‐tier approach to replicate stateful deterministic CORBA objects and allows object replicas to run on object request brokers from different vendors. A description of an IRL prototype developed in our department is proposed along with an extensive performance analysis. Copyright © 2003 John Wiley &amp; Sons, Ltd.

An architecture for highly available wide-area service composition

Service composition provides a flexible way to quickly enable new application functionalities in next generation networks. We focus on the scenario where next generation portal providers ‘compose’ the component services of other providers. We have developed an architecture based on an overlay network of service clusters to provide failure-resilient composition of services across the wide-area Internet: our algorithms detect and recover quickly from failures in composed client sessions. In this paper, we present an evaluation of our architecture whose overarching goal is quick recovery of client sessions. The evaluation of an Internet-scale system like ours is challenging. Simulations do not capture true workload conditions and Internet-wide deployments are often infeasible. We have developed an emulation platform for our evaluation—one that allows a realistic and controlled design study. Our experiments show that the control overhead involved in implementing our recovery mechanism is minimal in terms of network as well as processor resources; minimal additional provisioning is required for this. Failure recovery can be effected using alternate service replicas within about 1 s after failure detection on Internet paths. We collect trace data to show that failure detection itself can be tight on wide-area Internet paths—within about 1.8 s. Failure detection and recovery within these time bounds represents a significant improvement over existing Internet path recovery mechanisms that take several tens of seconds to a few minutes.

Client-access protocols for replicated services

Addresses the problem of replicated service provision in distributed systems. Existing systems that follow the state machine approach concentrate on the synchronization of the server replicas and do not consider the problem of client interaction with the server group. This paper analyzes client interaction and identifies a number of access protocols to meet a range of client requirements and system models. The paper demonstrates that protocols for the open group model-clients external to the group of servers-satisfy the requirements of the state machine approach, even when replication is transparent to the clients. Experimental performance results indicate that the open model is clearly desirable when the service is used by a large, dynamically changing set of clients, a situation which pertains to Internet service provision.

Evaluating the performance of flood-d: a tool for efficiently replicating Internet information services

Internet information services replicate their servers to improve availability, response time, and fault tolerance. Traditional replication algorithms do not address the scale and administrative decentralization of today's internetworks. We have proposed and implemented a scalable and efficient tool to replicate wide-area, autonomously managed services. We target replication degrees of thousands of weakly consistent replicas. The main goal of our replication tool is to make traditional replication services scale in today's exponentially growing, autonomously managed internetworks. Our tool, which we call flood-d, allows servers to be organized in multiple replication groups. For each replication group, flood-d builds a logical update topology that is resilient to server failure, and tries to minimize the communication cost and propagation time needed to transmit updates. Flood-d's logical topologies are computed based on communication latency and available network bandwidth. This paper describes flood-d, and presents simulation results obtained when using flood-d to extend existing replication algorithms. Our results show the gains of organizing service replicas into multiple, smaller replication groups, and using network-cognizant logical topologies to propagate updates. We argue that existing as well as emerging Internet information services can benefit from flood-d's services.

A fault tolerant RPC mechanism based on IP multicasting

This paper presents a fault tolerant Remote Procedure Call (RPC) mechanism based on Sun RPC and IP multicast protocol. The fault tolerant RPC mechanism is provided as an RPC library, called libFTRPC, whose interface is compatible with that of conventional Sun's RPC library. Thanks to this compatibility, a reliable RPC server can be developed in the same way a conventional RPC server is constructed. The service reliability is ensured by replicating the server to a group of server replicas. Coordinator-cohort replication model in conjunction with read-one/write-all policy is used to guarantee state consistency between server replicas. We also incorporate a simple and effective approach to balance loads among server replicas to improve performance for read-only requests.