Re: Oracle 10G

Obnoxio The Clown wrote: > Oh, for f***'s sake. This is extremely disingenious, even for you. IDS has > had this facility for years, there is nothing grid-y about it. Take the pills! Be careful... You're probably not young anymore >:> It has "hash" and "table" in it... it's distributed over several physical disks... Why not call it grid?! You are too strick! :) Regards.

Fernando Nunes wrote: > Obnoxio The Clown wrote: > >> Oh, for f***'s sake. This is extremely disingenious, even for you. IDS >> has had this facility for years, there is nothing grid-y about it. > > > Take the pills! Be careful... You're probably not young anymore >:> > It has "hash" and "table" in it... it's distributed over several > physical disks... Why not call it grid?! > You are too strick! :) > > Regards. I'm still waiting to here how one does transparent fail-over with DB2 or Informix. According to OTC you folks have had it for years and I have somehow missed that fact. -- 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)

Daniel Morgan wrote: > I'm still waiting to here how one does transparent fail-over with DB2 > or Informix. According to OTC you folks have had it for years and I "you folks" ??? Does that mean u are not part of us :-). Nothing wrong in being in Oracle camp, but at least don't have any pretentions of impartiality, which u brag about. I take this opportunity to post an old exchange between us. See whether it is still relevant:- =============================== From: rkusenet (rkusenet@sympatico.ca) Subject: Re: Fw: Company thought DB2 will be better than Oracle. View this article only Newsgroups: comp.databases.informix Date: 2003-09-23 07:09:01 PST Daniel Morgan wrote: >I think marketing, advertising, and hyperbole are irrelevant to a usenet group. >A bunch of employees crowing about the fact that the company that pays their >salary is better than the company that doesn't it not something I have to go to the > internet to read. Why restrict to only those who are employees of company that makes a product. In what way an Oracle user, specially DBAs, who uses Oracle for his living, will be any less biased than an employee of Oracle Corp like Mark Townsend. Of course the same applies to Informix DBAs too. > I've no doubt the people that are paid to do marketing > for Coca Cola can quote equally valid statistics about those >"Pepsi Users". Me? I just walk up to the machine, > put in my quarters and take whatever comes out. It's all brown fizzy water to me. the only difference is that coke/pepsi isn't about my job/career. Sure u and I will be unbiased when it comes to coke vs pepsi or Honda vs Toyota, since it is not about our job. But Oracle vs DB2/Informix is a different beast altogether. That's why I take ur statements defending Oracle's TCO with a handful of the needful. rk- > have somehow missed that fact. >

> an employee of Oracle Corp like Mark Townsend. Que ? Anyhow, to cut a long thread even longer, Oracle's definition of Grid Computing revolves around 4 key areas Storage Grid ------------ Storing all of your data in one or more Oracle databases on shared pools of disks, with automatic striping and mirroring across the pool to ensure maximum throughput and availability. Ability to add and remove disk to and from the pool on demand, and have the data automatically re-stripe and remirror online (to ensure max performance and availability). Built in data protection capabilities (mirroring, automated backup/recovery to a central recovery area, flashback (undo), and automated standby). Database Grid ------------- A cluster of standardized, low-cost commodity 2 and 4 way boxes, indirectly attached to above mentioned disk farms, running multiple defined workloads (i.e ERP, CRM, DW). Service level definitions built into the Oracle network and cluster level that detemines which nodes in the cluster are addressing which workload at which time, based on availability and performance service level definitions. Ability to dynamically adjust the amount of resources being dedicated at any one time to any one workload, based on resource consumption (Think of it as a cluster wide resource management framework). Supported on all OS's, without the need for OS specific clusterware. Databases are self managing, data stored in a database on one OS can be used by a database on a different OS. Ability to offload batch workload from one database to another via a built in scheduler that is integrated with the resource management, and ship not only the workload, but also the required data as well. Across OSes. With or without the cluster. App Server Grid ---------------- Where the more 'understood' parts of the traditional Grid lives - App Server farms (built from the same low cost hardware) supporting discovery, net-based communcation protocols, web services, application integration, GDK's etc. What Serge has been refering to. Note that the App Server layer shares the same workload definitions as the database layer, and is aware of where the underlying database resources are currently coming from for failover and resource balancing. Identity is also common across app server and database. Other than that it's all standards and alphabet soup to me and something I understand very little about. Grid Control ------------ Oracle management platform that spans the above. Provides end-to-end 'vertical' performance tracing and monitoring of a workload down through the horizontal layers of the 'Grid' (i.e from client machine, into network layer, into app server layer, into database layer, down into storage layer). Service level alerts built down through stack for performance and availability. Click to EJB and Click to SQL diagnosis for key representative transactions. Also provides central software configuration repository with built in patch management, supporting online checks with published reference configurations, and automated patch download. Centralized cloning and configfuration of new hardware resources as they are added to Grid. Built in policy violation checks of cloned environments. Allows new resources to be statically provisioned if the current dynamic resources are not enough. I think the big difference between Oracle's view of the Grid and what the Grid is currently known as (as least by IBM), is that most Grids today are attempted by orgs that typically cannot afford to provision their own IT environment, so they try to use the current Grid technology etc to find and 're-purpose' spare cycles on the net (or within a defined community). Sort of a computing socialism. The apps also tend to need to be 'embarassingly parallel' and specifically written to take advantage of the Grid like capabilities. A great idea for Universities doing large scale number crunching etc, and something that is supported in the App Server, but not too likely to work for Ford and GM for their business processing (Can you imagine Ford sending GM some of their SAP workload every now and then ?). Oracle's grid (at least for the data serving services below the app server) is based on the idea that companies will simply build and provision their own inexpensive 'grids' from low-cost components within their firewall, the so-called 'Enterprise Grid', and then run all of their applications on them transparently. Cost savings comes from standardization, automation, consolidation and commoditization. Top of line pay back comes from the ability to deploy new workloads quickly and dynamically share resources. Oh, and a big part of the Oracle thing is that it's also structured so that each bit provides some benefit in it's own right, nobody has to use all of it at once, it's OS and hardware idependent, and that existing customer investments are protected as well. So it's built on some things you have already seen (RAC, for instance), and adds some new capabilities as well.

rkusenet wrote: > Daniel Morgan wrote: > >> I'm still waiting to here how one does transparent fail-over with DB2 >> or Informix. According to OTC you folks have had it for years and I > > > "you folks" ??? That was sarcasm. The only year in the last 5 or 6 I haven't done some Informix work, so far, is this one. But I did have the joy of getting rid of a FoxPro database in January. That should count for something. -- 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)

Daniel Morgan wrote: > Fernando Nunes wrote: > >> Obnoxio The Clown wrote: >> >>> Oh, for f***'s sake. This is extremely disingenious, even for you. IDS >>> has had this facility for years, there is nothing grid-y about it. >> >> >> Take the pills! Be careful... You're probably not young anymore >:> >> It has "hash" and "table" in it... it's distributed over several >> physical disks... Why not call it grid?! >> You are too strick! :) >> >> Regards. > > I'm still waiting to here how one does transparent fail-over with DB2 > or Informix. According to OTC you folks have had it for years and I > have somehow missed that fact. What are you talking about? You talked about hash partitioned tables, not transparent failover. And many different types of partitioning have been in IDS for years. You then tried to imply some sort of grid-ness, because these might be on different vendors' disks.... -- "C'est pas parce qu'on n'a rien ''' dire qu'il faut fermer sa gueule" - Coluche