None

none

<p>To achieve better scalability, a fully replicated, distributed, main-memory database is divided into subparts, called segments. Segments may have individual degrees of redundancy and other properties that can be used for replication control. Segmentation is examined for the opportunity of decreasing replication effort, lower memory requirements and decrease node recovery times. Typical usage scenarios are distributed databases with many nodes where only a small number of the nodes share information. We present a framework for virtual full replication that implements segments with scheduled replication of updates between sharing nodes.</p><p>Selective replication control needs information about the application semantics that is specified using segment properties, which includes consistency classes and other properties. We define a syntax for specifying the application semantics and segment properties for the segmented database. In particular, properties of segments that are subject to hard real-time constraints must be specified. We also analyze the potential improvements for such an architecture.</p>

To achieve better scalability, a fully replicated, distributed, main-memory database is divided into subparts, called segments. Segments may have individual degrees of redundancy and other properties that can be used for replication control. Segmentation is examined for the opportunity of decreasing replication effort, lower memory requirements and decrease node recovery times. Typical usage scenarios are distributed databases with many nodes where only a small number of the nodes share information. We present a framework for virtual full replication that implements segments with scheduled replication of updates between sharing nodes. Selective replication control needs information about the application semantics that is specified using segment properties, which includes consistency classes and other properties. We define a syntax for specifying the application semantics and segment properties for the segmented database. In particular, properties of segments that are subject to hard real-time constraints must be specified. We also analyze the potential improvements for such an architecture.

<p>A distributed, real-time database could be used to implement a shared whiteboard architecture used for communication between mobile nodes, in an ad-hoc network. This kind of application implies specific requirements on how the database handles replication and consistency between replicas (global consistency). Since mobile nodes are likely to disconnect from the network and connect again at unpredictable times, and since a node may be disconnected an arbitrary amount of time, this needs to be treated as normal operation, and not as failures.</p><p>The replication scheme used in the DeeDS architecture, and the PRiDe replication protocol are both suitable for a shared whiteboard architecture as described above. Since the mobile nodes are likely to be some kind of hand-held device (e.g., used by rescue personnel to exchange information), the database system should be suitable for use in embedded systems. The Polyhedra Real-Time Relational Database (RTRDB) and the TimesTen database are two such systems. A problem is that neither of these two database systems have a replication scheme suitable for use in the previously described type of architecture.</p><p>This dissertation presents two design proposals for how to extend the Polyhedra RTRDB with support for multi-master replication of data using full replication and eventual consistency. One design proposal is based on the DeeDS architecture and the other is based on the PRiDe replication protocol. The proposal based on DeeDS puts a number of requirements on the underlying database and is not easy to port to another DBMS since it makes use of Polyhedra specific API’s. The proposal based on PRiDe on the other hand requires no instrumentation of the underlying database and is thus easier to port to other database systems.</p>

A distributed, real-time database could be used to implement a shared whiteboard architecture used for communication between mobile nodes, in an ad-hoc network. This kind of application implies specific requirements on how the database handles replication and consistency between replicas (global consistency). Since mobile nodes are likely to disconnect from the network and connect again at unpredictable times, and since a node may be disconnected an arbitrary amount of time, this needs to be treated as normal operation, and not as failures. The replication scheme used in the DeeDS architecture, and the PRiDe replication protocol are both suitable for a shared whiteboard architecture as described above. Since the mobile nodes are likely to be some kind of hand-held device (e.g., used by rescue personnel to exchange information), the database system should be suitable for use in embedded systems. The Polyhedra Real-Time Relational Database (RTRDB) and the TimesTen database are two such systems. A problem is that neither of these two database systems have a replication scheme suitable for use in the previously described type of architecture. This dissertation presents two design proposals for how to extend the Polyhedra RTRDB with support for multi-master replication of data using full replication and eventual consistency. One design proposal is based on the DeeDS architecture and the other is based on the PRiDe replication protocol. The proposal based on DeeDS puts a number of requirements on the underlying database and is not easy to port to another DBMS since it makes use of Polyhedra specific API’s. The proposal based on PRiDe on the other hand requires no instrumentation of the underlying database and is thus easier to port to other database systems.