Tuesday, 7 August 2007
What am (was) I doing in these projects?
My actual contribution on DBE was designing a distributed transaction model (it is reflected on DBE Report D24.5);
Web services are the fundamental elements Service-Oriented Computing (SOC) which emerges as the major paradigm for developing applications/solutions. They are loosely-coupled and can be composed to create more complex blocks of services. In the business terminology, these services must be performed in unison to accomplish a goal which is called transaction. The long-term nature of business transactions, in a highly dynamic and distributed environment, is the major challenge for constructing/defining a consistent transaction model. Disregarding the primary characteristics of SOC (ex. loose-coupling) or ignoring some important business requirements (ex. partial results), are important exertions in proposed transaction models, which also suffer from unnecessary complications of implementing a consistency model on top of the service-realisation boundary. In addition, the feasibility of a heavy coordinator framework misleads some transaction models to use a centralised (or limited decentralised) coordination model. Our model, by introducing a consistency graph for a transaction (IDG) tries to implement a distributed coordinator model in which a transaction can be performed through the cooperation (sharing of results) of the local coordinators. Furthermore, the possibility for realising partial results between different transactions is considered by introducing an External Dependency Graph, which in combination with the IDG can provide a global consistency model. On the localised design of the coordinator we introduce five new locks which not only provide consistency connection between the distributed logs (EDG and IDG) and the local concurrency control model but also cover dynamic aspects of the highly dynamic environment and isolate recovery management.
On OPAALS, our main goal is designing a Dynamic P2P architectures for distributed long-lived transactions:
Our research hypothesis is that in order to do this, we must move away from “traditional” centralised solutions for P2P architectures and transactional modelling, to fully distributed solutions. In keeping with the spirit of OPAALS our P2P architecture proposals derive much of their inspiration from studies of networks of interaction and collaboration in the social and natural sciences. In particular the P2P network model can be viewed as an example of a birth (node duplication with divergence), death (node inactivation or removal), and innovation (horizontal “gene” transfer) model (BDIM). What we are aiming for here, is the development of a digital environment that evolves naturally to support a sustainable economy, where interactions through e-commerce are an integral part of that economy…
Some references:
Web services are the fundamental elements Service-Oriented Computing (SOC) which emerges as the major paradigm for developing applications/solutions. They are loosely-coupled and can be composed to create more complex blocks of services. In the business terminology, these services must be performed in unison to accomplish a goal which is called transaction. The long-term nature of business transactions, in a highly dynamic and distributed environment, is the major challenge for constructing/defining a consistent transaction model. Disregarding the primary characteristics of SOC (ex. loose-coupling) or ignoring some important business requirements (ex. partial results), are important exertions in proposed transaction models, which also suffer from unnecessary complications of implementing a consistency model on top of the service-realisation boundary. In addition, the feasibility of a heavy coordinator framework misleads some transaction models to use a centralised (or limited decentralised) coordination model. Our model, by introducing a consistency graph for a transaction (IDG) tries to implement a distributed coordinator model in which a transaction can be performed through the cooperation (sharing of results) of the local coordinators. Furthermore, the possibility for realising partial results between different transactions is considered by introducing an External Dependency Graph, which in combination with the IDG can provide a global consistency model. On the localised design of the coordinator we introduce five new locks which not only provide consistency connection between the distributed logs (EDG and IDG) and the local concurrency control model but also cover dynamic aspects of the highly dynamic environment and isolate recovery management.
On OPAALS, our main goal is designing a Dynamic P2P architectures for distributed long-lived transactions:
Our research hypothesis is that in order to do this, we must move away from “traditional” centralised solutions for P2P architectures and transactional modelling, to fully distributed solutions. In keeping with the spirit of OPAALS our P2P architecture proposals derive much of their inspiration from studies of networks of interaction and collaboration in the social and natural sciences. In particular the P2P network model can be viewed as an example of a birth (node duplication with divergence), death (node inactivation or removal), and innovation (horizontal “gene” transfer) model (BDIM). What we are aiming for here, is the development of a digital environment that evolves naturally to support a sustainable economy, where interactions through e-commerce are an integral part of that economy…
Some references:
- A. Razavi, S. Moschoyiannis and P. Krause. Concurrency Control and Recovery Management for Open e-Business Transactions. In Proc. of Communicating Process Architectures (CPA 2007), 2007.
- A. Razavi, S. Moschoyiannis and P. Krause. A Coordination Model for Distributed Transactions in Digital Business Ecosystems. In Proc. of IEEE Int'l Conf. on Digital Ecosystems and Technologies (IEEE-DEST 2007), IEEE Computer Society, 2007. [preliminary version: PDF]
- A. Razavi, P. Malone, S. Moschoyiannis, B. Jennings and P. Krause. A Distributed Transaction and Accounting Model for Digital Ecosystem Composed Services. In Proc. of IEEE Int'l Conf. on Digital Ecosystems and Technologies (IEEE-DEST 2007), IEEE Computer Society, 2007. [preliminary version: PDF]
- A. Razavi, S. Moschoyiannis and P. J. Krause Preliminary Architecture for Autopoietic P2P Network focusing on Hierarchical Super-Peers, Birth and Growth Models. OPAALS project Deliverable D3.1, 2007
- A. R. Razavi, P. J .Krause and S. K. Moschoyiannis. DBE Report D24.5: DBE Distributed Transaction Model, University of Surrey, 2006.
Labels: What I am doing here
// posted by Lost in the crowd! @ 09:11 
Subscribe to Posts [Atom]
