Re: Oracle 10G

Daniel Morgan <damorgan@x.washington.edu> wrote in message news:<1078122180.285284@yasure>... > Distributed hash tables aren't in a commercial DBMS yet? > > http://download-west.oracle.com/docs/cd/B10501_01/server.920/a96540/statements_73a.htm#2063017 > > CREATE TABLE hash_products > ( product_id NUMBER(6) [ snip ] > ) ) > PARTITION BY HASH (product_id) > PARTITIONS 5 > STORE IN (tbs_1, tbs_2, tbs_3, tbs_4); > > Could have fooled me. Your choice of words. You teach a university course on data management, and you don't even know the terms of art? A distributed hash table (or distributed hashing) is a technique for allocating keyed data values over a set of machines. In contrast to the hash partitioning you have presented here, in a distributed hash each storage resource is autonomous and posesses limited knowledge about its peers, and no single machine knows where every partition is. Data lookup is accomplished by a variety of message forwarding techniques. A grid is a physical architecture. In a grid, machines that participate are autonomous peers. Their resources -- storage, compute, memory, etc -- are 'virtualized' with the grid software performing the physical management. > All I care about is the technology. And no matter what Oracle > chooses to call it ... it is incredible. [snarky mode on] To judge from your lack of familiarity with technical language and what it means, I am skeptical of your judgements. It sounds to me like O have, in 10G, on-line provisioning. Great. It's hard to do. But others -- Teradata, Digital Rdb, Tandem and mainframe DBMSs, to pick some examples -- have had this for years (decades, in the case of Rdb and mainframe DBMSs). Pb

Paul G. Brown wrote: > You teach a university course on data management, and you don't even > know the terms of art? We aren't discussing my "academic" definitions of these words. We are discussing those definitions as developed by marketing departments. There about as many definitions of "grid" as their are commercial implementations. Your definition of distributed hashing as only being of relevance to shared-nothing architecture is nothing but the same. It is the definition based upon a specific architecture: It is not the definition based upon all architectures. -- Daniel Morgan http://www.outreach.washington.edu/ext/certificates/oad/oad_crs.asp http://www.outreach.washington.edu/ext/certificates/aoa/aoa_crs.asp damorgan@x.washington.edu (replace 'x' with a 'u' to reply)