Sebnem Bora

Applying feedback control in adaptive replication mechanisms in fault tolerant multi-agent organizations

Effective fault-handling in emerging complex applications in large-scale MAS (Multi-agent Systems) requires the ability to dynamically adapt resource allocation and fault tolerance policies in response to changes in environment, user or application requirements, and available resources. This adaptation process incorporates an observation mechanism that transparently monitors the applications behaviors as well as the availability of resources, and adaptively reconfigures the system resources. This process is realized by a specific module which exploits the information resulting from monitoring. In this paper, we present an approach for adaptive replication. This approach uses an observation mechanism and a feedback control system within an adaptive replication infrastructure to support adaptive fault tolerance in multi-agent organizations. The main strategy used in our approach is to insert control theory methodology and analysis to adaptive replication. Thus, our approach provides a systematic and scientific method for implementing adaptive fault tolerance policies in MAS.

Implementing a multi-agent organization that changes its fault tolerance policy at run-time

In this paper, we present an approach that supports simultaneously applying different fault tolerance policies in multi-agent organizations. The main strategy of our approach is to implement fault tolerance policies as reusable agent plans using HTN (Hierarchical Task Network) formalism. In this way, different fault tolerance policies such as static and adaptive ones can be implemented as different plans. In a static fault tolerance policy, all parameters related to the fault tolerance are set by a programmer before run-time. However, an adaptive fault tolerance policy requires dynamically adapting resource allocation and replication mechanisms by monitoring the system. Monitoring of a system brings some cost to the system. If all agents in an organization apply the adaptive fault tolerance policy, the monitoring cost will become an important factor for the system performance. Hence by applying our approach, the adaptive policy can be applied only to the critical agents whose criticalities can be observed during the organizations lifetime and the static one can be applied to the remaining agents. This reduces the monitoring cost and increases the overall organization performance. A case study has been implemented to show the effectiveness of our approach.

A Performance and Profit Oriented Data Replication Strategy for Cloud Systems

In todays world, tenants of cloud systems expect timely responses to queries that process ever-increasing sizes of data. However, most cloud providers offer their services without any performance guarantees to their tenants. In this paper we propose a data replication strategy that aims to satisfy performance guarantees for the tenant while ensuring profitability of the cloud provider. Our strategy estimates the response time of the queries, as well as the expenditures that affect the profitability of the cloud provider. The decision of whether to perform replication is determined by the fulfillment of these two criteria. Validity of the proposed strategy is provided by means of a simulation study.

A fault tolerant system using collaborative agents

Replication of data or processes is an effective way to provide enhanced performance, high availability and fault tolerance in distributed systems. For instance, in systems based on the client-server model, a server may serve many clients and because of heavy loads, the server cannot respond to the requests on time. In such a case, replicating data or servers may improve performance. Moreover, data and processes can be replicated to protect against failures. However, this is a very complex procedure. In this paper, I propose a method, to make systems fault tolerant based on replication, by way of exploiting the use of collaborative agents. This method is also used to improve fault tolerance in multi-agent systems.

Ensuring performance and provider profit through data replication in cloud systems

Cloud computing is a relatively recent computing paradigm that is often the answer for dealing with large amounts of data. Tenants expect the cloud providers to keep supplying an agreed upon quality of service, while cloud providers aim to increase profits as it is a key ingredient of any economic enterprise. In this paper, we propose a data replication strategy for cloud systems that satisfies the response time objective for executing queries while simultaneously enables the provider to return a profit from each execution. The proposed strategy estimates the response time of the queries and performs data replication in a way that the execution of any particular query is still estimated to be profitable for the provider. We show with simulations that how the proposed strategy fulfills these two criteria.

On the choice of sampling rates in a fault-tolerant multi-agent system

Adaptive replication increases the systems response time due to the need for monitoring in fault tolerant multi-agent systems. The sampling period is one of the key factors that could influence the cost of adaptive replication. In this paper, we show how to select an appropriate sampling period in a heuristic manner to decrease the cost incurred by adaptive replication.

Experience with Feedback Control Mechanisms in Self-replicating Multi-Agent Systems

In this paper, we present an approach for adaptive replication to support fault tolerance. This approach uses a feedback control theory methodology within an adaptive replication infrastructure to determine replication degrees of replica groups. We implemented this approach in a multi-agent system to survive Byzantine failures. At the end of the paper, we also provide some experimental results to show the effectiveness of our approach.