Chamberlin, D., Schmuck, F. Dynamic Data Distribution (D 3 ) in a Shared-Nothing Multiprocessor Data Store. VLDB-92, 1992.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....Another direction is to devise variants for high availability in the presence of presplitting. The spare creation algorithm for these variants should be revisited. Another direction for future work is to devise schemes for higher throughput through replication using more than one mirror, as in [ChS92]. Such schemes are typically too expensive as a solution to the sole 12 06 12 2000 11:37 problem of bucket failures. However, they are useful for geographically distributed applications with local sites where they increase the throughput for searches, as long as the network does not become ....

Chamberlin, D., Schmuck, F. Dynamic Data Distribution (D 3 ) in a Shared-Nothing Multiprocessor Data Store. VLDB-92, 1992.

....horizontal scalability over commodity components, e.g. mass produced Wintel or Unix boxes like RS6000, Alpha etc. and networks like Fast Ethernet or ATM, A al95] I98] M96] M97] M97a] Large and scalable data collections critically demand high availability and security of the stored data, [ChS92], B al95] H96] T95a] L96] A al97] Traditional high availability storage techniques provide for limited horizontal scalability, including the hardware RAID schemes, PGK88] R97] KH98] and, more recent software RAID schemes, e.g. within Windows NT and others [SS90] MRW95] RM96] ....

....and eighties. Their goal was however high search performance rather than high availability. Their prime disadvantage is the high storage cost. This limits the approach in practice to files providing 1 availability at best. The application of mirroring to scalable multicomputer files was studied in [ChS92], and, through the LH m schema in [LN96a] for LH files specifically. M P k J 4 0.957 32 1 8 0.957 32 1 16 0.957 32 1 32 0.957 32 1 64 0.952 16 1 128 0.946 8 1 256 0.963 32 2 512 0.977 16 2 1024 0.954 16 2 2048 0.971 8 2 4096 0.980 4 2 8192 0.961 4 2 16384 0.965 16 3 32768 0.987 ....

Chamberlin, D., Schmuck, F. Dynamic Data Distribution (D 3 ) in a Shared-Nothing Multiprocessor Data Store. VLDB-92, 1992.

....concern the high availability SDDSs, able to serve any data stored despite unavailability (failure, inaccessibility. of some storage sites (nodes) Research showed that direct adaptation of known recipes for high availability to an SDDS may not be practical. Mirroring may cost too much [ChS92], LN96a] The protection against unavailability of even a single storage site of an SDDS requires doubling the number of storage sites; a substantial cost for a larger multicomputer, e.g. the 166 site multicomputer of Inktomi at Santa Clara, CA, or the 100site one of Yahoo in Vienna, VA. More ....

Chamberlin, D., Schmuck, F. Dynamic Data Distribution (D 3 ) in a Shared-Nothing Multiprocessor Data Store. VLDB-92, 1992.

....either getting slowly out of date (e.g. books lists) or bringing few considerable incoherence (e.g. statistical studies data) Many studies have been carried out in the context of OLTP or QP. The main topics are data fragmentation [1,2] Parallel Execution Plans [3,4] and duplication strategies [5]. Most of the work done was built on the assumption that parallel DBMS would run on Massively Parallel Machines (MPM) Such machines, while causing an incomparable rise of performance, are still quite few, and represent a big investment. This made hybrid architectures, such as workstation ....

D. Chamberlin and F. Shmuck. Dynamic Data Distribution (D3) in a Shared-Nothing Multiprocessor Data Store. In Proceedings of the 18 th VLDB Conference, Vancouver, British Columbia, Canada, 1992.

....Speed of updates, and time in general, are less significant here, as insconsistencies appear slower, and are generally less dangerous. Many studies have been carried out in these contexts. The main topics are data fragmentation [1,2] Parallel Execution Plans [3,4] and duplication strategies [5]. Most of the work done was designed for Massively Parallel Machines (MPM) which are powerful but quite expensive. This made hybrid architectures, such as workstation clusters, or networks of workstations come to the front page of research [6,7] More than suggestive aspects, such as global cost, ....

D. Chamberlin and F. Schmuck, "Dynamic Data Distribution (D3) in a SharedNothing Multiprocessor Data Store," in Proceedings of the 18th VLDB Conference, (Vancouver, British Columbia, Canada), 1992.

....another at any time a certain imbalance in load is detected. Migration of entire files has been considered in the context of replicated file systems. On the other hand, migration of file portions has been considered for scalable, distributed hashing schemes but with different objective functions [2, 6, 9, 50, 75, 76, 83]. The only work that considers data migration in the context of disk load balancing is [38] however, this work is restricted to offline and monolithic (i.e. non incremental) reorganization. The load balancing component of our intelligent file system consists of two independent modules: one that ....

Chamberlin DD, Schmuck FB (1992) Dynamic Data Distribution (D<F3.733e+05> 3<F3.733e+05> ) in a Shared-Nothing Multiprocessor Data Store. International Conference on Very Large Data Bases, Vancouver, pp 163--174

No context found.

Chamberlin, D., Schmuck, F. Dynamic Data Distribution (D 3 ) in a Shared-Nothing Multiprocessor Data Store. VLDB-92, 1992.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC