\input zb-basic
\input zb-ioport
\iteman{io-port 01935524}
\itemau{Liu, Xiangning; Helal, Abdelsalam; Du, Weimin}
\itemti{Multiview access protocols for large-scale replication.}
\itemso{ACM Trans. Database Syst. 23, No. 2, 158-198 (1998).}
\itemab
Summary: \Beginpar{}{}{}The article proposes a scalable protocol for replication management in large-scale replicated systems. The protocol organizes sites and data replicas into a tree-structured, hierarchical cluster architecture. The basic idea of the protocol is to accomplish the complex task of updating replicated data with a very large number of replicas by a set of related but independently committed transactions. Each transaction is responsible for updating replicas in exactly one cluster and invoking additional transactions for member clusters. Primary copies (one from each cluster) are updated by a cross-cluster transaction. Then each cluster is independently updated by a separate transaction. This decoupled update propagation process results in possible multiple views of replicated data in a cluster. Compared to other replicated data management protocols, the proposed protocol has several unique advantages. First, thanks to a smaller number of replicas each transaction needs to atomically update in a cluster, the protocol significantly reduces the transaction abort rate, which tends to soar in large transactional systems. Second, the protocol improves user-level transaction response time as top-level update transactions are allowed to commit before all replicas have been updated. Third, read-only queries have the flexibility to see database views of different degrees of consistency and data currency. This ranges from global, most up to date, and consistent views, to local, consistent, but potentially old views, to local, nearest to users but potentially inconsistent views. Fourth, the protocol maintains its scalability by allowing dynamic system reconfiguration as it grows by splitting a cluster into two or more smaller ones. Fifth, autonomy of the clusters is preserved as no specific protocol is required to update replicas within the same cluster. Clusters are, therefore, free to use any valid replication or concurrency control protocols.\Endpar (Provider: ACM) Review: \Beginpar{}{}{}The authors propose a scalable protocol for managing replicated data in large-scale systems. Extensions to previous replication management protocols are predicated on the need to handle data copies at hundreds and even thousands of locations. The multiview access (MVA) protocol organizes sites and data copies into a tree-structured, hierarchical cluster architecture. The key to the MVA protocol is the ability to decompose a large update transaction into a related set of independently committed update transactions. Each update transaction is responsible for updating all replicas in one cluster and the primary copy in all of its child clusters. Then new update transactions are invoked in all the child clusters. Beginning at the root cluster, this protocol continues until all data copies are updated. Thus, at any time, multiple views of replicated data can appear in the system. Correctness proofs are provided for the basic protocol and for an extended protocol with backup primary and coordinator sites.\Endpar \Beginpar{}{}{}Five unique advantages are claimed for the MVA protocol: \SGMPbeginList{}{bulleted}{}{}{}{} \SGMPbeginItem \Beginpar{}{}{}Due to the smaller number of copies to be updated, the transaction abort rate is reduced.\Endpar \SGMPendItem \SGMPbeginItem \Beginpar{}{}{}Transaction response times are improved as top-level update transactions are allowed to commit before all copies have been updated.\Endpar \SGMPendItem \SGMPbeginItem \Beginpar{}{}{}Read-only queries have the flexibility to request a desired level of data consistency and currency from the multiple database views.\Endpar \SGMPendItem \SGMPbeginItem \Beginpar{}{}{}The MVA protocol is scalable, since clusters can be dynamically split or combined.\Endpar \SGMPendItem \SGMPbeginItem \Beginpar{}{}{}Autonomy of the clusters is preserved since no specific protocol is required to update copies within the cluster.\Endpar \SGMPendItem \SGMPendList The performance claims are validated via simulation experiments.\Endpar \Beginpar{}{}{}The MVA protocol is an interesting and seemingly practical approach for providing replication management in large-scale systems, such as large organizational intranets or even the Internet. More research and evaluation will be needed before actual implementation, but developers of large distributed systems with significant data replication requirements are encouraged to read this paper and understand the basic concepts of the proposed approach.\Endpar (Provider: ACM)
\itemrv{~}
\itemcc{H.2.4 H.2.4 C.2.4}
\itemut{algorithms; design; experimentation; measurement; management; performance}
\itemli{doi:10.1145/292481.277628}
\end