Scalable Distributed Data Structures Research Articles

Fault-tolerant control for Scalable Distributed Data Structures

calable Distributed Data Structures (SDDS) can be applied for multicomputers. Multicomputers were developed as a response to market demand for scalable and dependable but not expensive systems. SDDS consists of two components dynamically spread across a multicomputer: records belonging to a file and a mechanism controlling record placement in the file. Methods of making records of the file more or less fault-tolerant have already been given. Methods of making the mechanism controlling record placement in the file fault-tolerant have not been studied, yet, although it seems that this is more important for the system dependability than record fault tolerance. Faults in control may lead an application to crash, while record data faults may cause invalid computations at most. In the paper a fault-tolerant control for SDDS is given. It is based on an application of Job Comparison Technique along with TMR. Time overhead due to redundancy introduced is estimated, too.

Using Actors to Build a Parallel DBMS

In this paper, we present the design and the architecture of a parallel main memory database management system. We focus on concurrency control scheme and recovery. Our prototype is based on the concept of "database actors", an object-oriented data model well suited for parallelmanipulations. The storage sub system is built upon distributed Ram-files using SDDS (Scalable Distributed Data Structures) techniques. A nested transaction model is proposed and used to handle concurrency access and recovery. We have also proposed novel approach, based on wait-die, to implement a distributed deadlock prevention technique for our model of nested transactions.

SDDSfL vs. local disk - a comparative study for Linux

Efficient data management and distribution in a multicomputer is a subject of much research. Distributed file systems are the most common solution of this problem, however, recent works are focused on more general data distribution protocols. Scalable, Distributed Data Structures (SDDS) are another promising approach to this issue. In this paper we discuss the efficiency of an implementation of SDDS in various applications. The results of experiments are presented.

Maintaining and checking parity in highly available Scalable Distributed Data Structures

Access to data stored in distributed main memory is much faster than access to local disks. Highly available, Scalable Distributed Data Structures (SDDS) utilize this fast access. They counteract the effects of failed or unavailable nodes by storing data redundantly. Since main memory per node is limited, they generate this redundancy by storing parity data calculated with erasure correcting codes instead of using replication. We present here a way to maintain parity that is about 10 times faster than using the traditional 2PC scheme. We also present a scheme that can diagnose a mismatch between parity and user data with very little network traffic.

Data warehouse access using multi-agent system

The new approach that we will propose, in this paper deals with the dynamic data distribution of the data warehouse (DWH) on a set of servers. This distribution is different from the "classical" one which depends on how data is used. It consists in distributing data when the machine reaches its storage limit capacity. The proposed approach insures the scalability and exploits the storage and processing resources available in the organization using the DWH. It is worth noting that our approach is based on a multi-agent model mixed with the scalability distribution proposed by the Scalable Distributed Data Structures. Our multi-agent model is made up of stationary agent classes: Client, Dispatcher, Domain and Server, and a mobile agent class: Messenger. These agents collaborate and achieve automatically the storage, splitting, redirection and access operations on the distributed DWH. In this paper, we focus on the global dynamic for the data access operation and we present the inherent experimental results.

AGENT BASED DATA STORAGE AND DISTRIBUTION IN DATA WAREHOUSES

The data warehouse (DWH) is usually presented as a centralized database. In this paper, we propose a new approach to manage data storage and distribution in a data warehouse environment. This approach deals with the dynamic data distribution of the DWH on a set of servers. The data distribution that we consider is different from the "classical" one which depends on the data use. The distribution in our approach consists in distributing data when the server reaches its storage capacity limit. This distribution assures the scalability and exploits the storage and processing resources available in the organization using the data warehouse. It is worth noting that our approach is based on a multi-agent model mixed with the scalability distribution proposed by the Scalable Distributed Data Structures. The proposed multi-agent model is composed of stationary agent classes: Client, Dispatcher, Domain and Server, and a mobile agent class called Messenger. These agents collaborate and interact to achieve automatically the storage, the splitting (distribution), the redirection and the access operations on the distributed data warehouse. In this paper, we demonstrate the improvements obtained when we have used the multi-agent system and the Messenger agents in the data storage operation.

Scalable distributed compact trie hashing (CTH*)

This last decade, a new class of data structures named Scalable Distributed Data Structures (SDDSs), is appeared completely dedicated to a distributed environment. This type of data structures opened an important axis of research, considering that the data management in a transparent manner is fundamental in a computer network. All the existing methods are mainly based on Linear hashing (LH*) and Range-partitioning (RP*). In this paper, we propose a new method with the constraints of the SDDS. Our approach is an adaptation of the well-known method Trie hashing (TH) for a distributed environment, i.e. a network of interconnected computers. The latter uses a digital tree (trie) as access function. Our major objective is the distribution of file buckets and the tree representing the hashing function. We have considered TH with the tree represented in compact form (CTH) because this option is probably more interesting for the reduction of the message size circulating on the network. Contrary to the majority of the existing methods, the proposed one provides the order of distributed files, then facilitates both the range query operations and the ordered traversal of files. Moreover, the following properties make our method a promising opening towards a new class of SDDS: (a) preservation of the order of records, (b) works without multicast (c) three bytes are sufficient to address a server, (d) the transfer of some bytes is enough for the update of the client trees. The access performances should exceed the ones of traditional files and some competitive scalable and distributed data structures.