1 INFO-VAX	Sat, 07 Aug 2004	Volume 2004 : Issue 434       Contents: Re: HSD partitioning Re: HSD partitioning# Re: HSD partitioning, fragmentation  Re: Phantom telnet connections Re: rx2600 and SAS Re: rx2600 and SAS Re: rx2600 and SAS Re: rx2600 and SAS RE: rx2600 and SAS Re: Solaris to Itanium... C Re: TECO manual location, was: Re: [Somewhat OT] What your non-OVMS C Re: TECO manual location, was: Re: [Somewhat OT] What your non-OVMS  Re: The free lunch YOU pay for3 Re: Time to patch OpenVMS - DCE-RPC buffer overflow  Re: Touch Screen Voting systems = Re: Ucx v4.0, network_mask set to 255.255.252.0 doesn't save. = Re: Ucx v4.0, network_mask set to 255.255.252.0 doesn't save. < [OT]: $1 million prize for Ingres ISV's - How about VMS too?  F ----------------------------------------------------------------------   Date: 6 Aug 2004 11:42:25 -0700 1 From: keithparris_NOSPAM@yahoo.com (Keith Parris)  Subject: Re: HSD partitioning = Message-ID: <cf15391e.0408061042.72cb6e04@posting.google.com>   l jox@unforgettable.com (john oxley) wrote in message news:<17ac92e.0408060048.63eddc93@posting.google.com>...F > Anyone know a way of splitting/partitioning a large disk on a HSD to+ > serve more than one unit to a VMS system?   2 What model of HSD controller are you working with?  ? I did a DECUS session on this issue a while back: see the "Disk # Partitioning on OpenVMS" pointer at % http://www.geocities.com/keithparris/   D Nigel Barker said "Host-based RAID-5" and that is certainly possible> with the "HP RAID Software for OpenVMS" softare product he was speaking of (see the SPD at D http://h18000.www1.hp.com/info/SP4649/SP4649PF.PDF) but you can alsoA do a host-based RAID-0 array, or even partition a single disk (by > forming a 1-member RAID-0 array, technically), using that same	 software.    ------------------------------  % Date: Fri, 06 Aug 2004 23:51:25 +0200  From: Dirk Munk <munk@home.nl> Subject: Re: HSD partitioning 2 Message-ID: <cf0ugu$61v$1@news1.tilbu1.nb.home.nl>   john oxley wrote: F > Anyone know a way of splitting/partitioning a large disk on a HSD to+ > serve more than one unit to a VMS system?   What kind of HSD? 99xx or 95xx ?   ------------------------------  % Date: Fri, 06 Aug 2004 18:55:06 -0400 " From: Glenn Everhart <gce@gce.com>, Subject: Re: HSD partitioning, fragmentation3 Message-ID: <41140cae$0$5919$61fed72c@news.rcn.com>   
 chrisl wrote: , >  john oxley  jox@unforgettable.com says... > F >>Anyone know a way of splitting/partitioning a large disk on a HSD to+ >>serve more than one unit to a VMS system?  >> >  > > > IIRC there's a CREATE_PARTITION command in the HSD software. > " Use vddriver or possible lddriver.  M Actually much depends on intended use. If the idea is to have partitions that K things like MSdos or unix or the like recognize, you may have to use one of L their tools, and then something like VDdriver pointing at the LBNs where theJ area(s) you want for VMS are to get a disk that can be used for a VMS fileK structure. The VMS boot code has no clue about partition tables. Those were H invented long ago in unix land for drivers that couldn't handle disks as big as existed...   F On another issue, the Fragmentation Avoider still works on VMS, thoughD it should be compiled from source to avoid symbol resolution issues.K It has been on sigtapes a few years back. Look for something like FA.zip or  favoid.zip.    ------------------------------  # Date: Fri, 06 Aug 2004 21:58:03 GMT 6 From: "Andy Bustamante" <a_c_bustamante@earthlink.net>' Subject: Re: Phantom telnet connections ; Message-ID: <L9TQc.2489$A_1.504@newssvr29.news.prodigy.com>   L Our solution was to upgrade to TCPIP 5.1 with ECOs.  For VMS 7.2-1h1 you canH apply the latest 5.1 ECO.  I don't know if Cisco was ever addressed thisK behavior.  We tried time outs and worked things from the firewall and Alpha J sides.  It's been a while since this was problem, since we're at TCPIP 5.3J now, but I know other people have seen simliar behavior with Pix firwalls.  I If you leave an idle connection (telnet or application telnet in my case) L open from an external network, occassionally, the client will time out.  TheK Pix will continue to hold the internal connection open.  This may result in F users showing logged in or my just hold BG devices open.  It was neverI consistent, but can be demontrated with enough external clients and a few  days effort.  L If your production environment requirements won't allow a TCPIP upgrade thenB you should start checking the firewall firmware and work this as a8 networking issue.  Cisco may have addressed this by now.     --       Andy Bustamante  Remove the ASCII 95s for e-mail     5 "Jack Peacock" <peacock@simconv.com> wrote in message , news:r4qdnc3nr5HuMI7cRVn-vQ@mpowercom.net...C > "Andy Bustamante" <a_c_bustamante@earthlink.net> wrote in message 5 > news:SXyQc.107$QJ3.71@newssvr21.news.prodigy.com... G > > I believe you're up to date for 5.0a.  We had some telnet issues at  sites I > > where Pix firewalls are in use with 5.1 corrected under the the later  > ECOs. K > > If upgrading TCPIP is an issue, is there a firewall holding connections I > > open?  Our problem was having telnet sessions time out to the client,  but L > > having the earlier Pix firmware hold the connection open on the internal > > network segment. > > F > The site does have a Pix firewall, and it is possible the hanging BG devices G > might be from infrequent telnet through the Pix building up over time  (last > > reboot 4 months ago).  Do you have any more details on this? >   Jack Peacock >  >    ------------------------------  # Date: Fri, 06 Aug 2004 20:44:07 GMT L From: winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr") Subject: Re: rx2600 and SAS 6 Message-ID: <00A35F43.E8CCCCB9@SSRL.SLAC.STANFORD.EDU>  _ In article <2nhmauFv0qlU1@uni-berlin.de>, Patrick Young <patrick@hilux.ace.unsw.EDU.AU> writes:  >Larry Kilgallen wrote: _ >> In article <trj4h0t4lt80o19p3rc8eghsaa2tfpe6mg@4ax.com>, Nigel Barker <nigel@hp.com> writes:  >>  J >>>On Thu, 5 Aug 2004 07:15:43 -0700, "Tom Linden" <tom@kednos.com> wrote: >>> G >>>>< There is no evaluation copy to download but you can purchase for   >>>>< $75 a SoftwareA >>>>< Developers Kit with all compilers & other layered products.  >>>># >>>>All? Well, with some compilers.  >>> S >>>OK, it ships with C, Fortran, COBOL & Macro32 plus you can download C++, BASIC & R >>>Pascal as they weren't available when the CDs were mastered. That's pretty muchL >>>all the compilers. Don't tell me you only wanted PL/1 as that isn't being
 >>>ported:-).  >>  G >> I gather it also does not include the Ada95 compiler VMS Development  >> said would be available.  > I >Many thanks to all who responded either by email or NG. It appears there G >might be a download available, but we are happy to purchase the kit if 
 >in stock. > G >Interesting, and pleased to note BASIC is available. IIRC VAX -> Alpha - >there was a consideration in dropping it :-(   H My understanding - unofficial, possibly incorrect - was that they didn'tJ consider dropping it; they forgot to include it in porting plans, realized their error somewhat later.      -- Alan      --  O =============================================================================== 0  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUM  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056 M  Paper mail to: SSRL -- SLAC BIN 99, 2575 Sand Hill Rd, Menlo Park CA   94025 O ===============================================================================    ------------------------------  % Date: Sat, 07 Aug 2004 04:59:20 +0800 , From: Paul Repacholi <prep@prep.synonet.com> Subject: Re: rx2600 and SAS 0 Message-ID: <87brhohtjr.fsf@k9.prep.synonet.com>  5 Patrick Young <patrick@hilux.ace.unsw.EDU.AU> writes:   ? > Cobol? - we have an application still written in it would you 
 > believe?  E A bank as part of the total rewrite they did for Y2K dropped PL/I and 0 did it ALL in COBOL. Not a line of C to be seen.   --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------   Date: 7 Aug 2004 00:24:08 GMT , From: bill@gw5.cs.uofs.edu (Bill Gunshannon) Subject: Re: rx2600 and SAS * Message-ID: <2nilp7F1batqU1@uni-berlin.de>  0 In article <87brhohtjr.fsf@k9.prep.synonet.com>,/ 	Paul Repacholi <prep@prep.synonet.com> writes: 7 > Patrick Young <patrick@hilux.ace.unsw.EDU.AU> writes:  > @ >> Cobol? - we have an application still written in it would you >> believe?  > G > A bank as part of the total rewrite they did for Y2K dropped PL/I and 2 > did it ALL in COBOL. Not a line of C to be seen.   E I was very happy as a COBOL programmer many years ago.  After reading E the articles posted here recently, maybe I should consider going back E to it.  It was the only programming I did on IBM mainframes (370VM on D one of the earliest 4331's) that I did not find at all painful.  ButB it was still like hitting yourself with a hammer.  Sure feels goodB when you stop.  I wnet on from there to UNIVAC COBOL with databaseE access and later Pr1me and the PDP-11.  Sadly I had left COBOL behind + before having the luck to discover the VAX.   7 I wonder where one looks to find real work doing COBOL?    bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------   Date: 6 Aug 2004 20:33:58 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen)  Subject: Re: rx2600 and SAS 3 Message-ID: <4IzmG3GFAGtG@eisner.encompasserve.org>   m In article <f401eb7f.0408060828.5bed43fc@posting.google.com>, soccer13player@yahoo.com (Nom de Plume) writes:  > Bill Gunshannon wrote:B >> So.....  As has been stated here the last time this subject wasA >> discussed, all the DOD applications that were certified under  G >> ADA on VMS were not certified with GNAT.  If they have to be re-done E >> and re-certified anyway, why would anyone choose to do it using an E >> inferior product like GNAT just to stay on VMS?  So there goes yet  >> another class of customers. > F > Having worked with many of the DEC compilers, I have found them very+ > nice.  How bad is GNAT compared to these?   D I have not used GNAT, but I distinctly remember the president of ACTD saying they felt that having a compiler generate listing files was a- mechanism of the past and no longer required.   C Since many of us rely on the listings VMS Development produces, the B example of retaining a record of released version is quite evidentA in VMS circles.  I fear that GNAT would be a considerable culture  shift.   ------------------------------  $ Date: Fri, 6 Aug 2004 20:13:17 -0700# From: "Tom Linden" <tom@kednos.com>  Subject: RE: rx2600 and SAS 9 Message-ID: <NDEMLKKEBOIFBMJLCECIEEDKDKAA.tom@kednos.com>    < -----Original Message-----5 < From: Paul Repacholi [mailto:prep@prep.synonet.com] ' < Sent: Friday, August 06, 2004 1:59 PM  < To: Info-VAX@Mvb.Saic.Com  < Subject: Re: rx2600 and SAS  <  < 7 < Patrick Young <patrick@hilux.ace.unsw.EDU.AU> writes:  < A < > Cobol? - we have an application still written in it would you  < > believe? < G < A bank as part of the total rewrite they did for Y2K dropped PL/I and 2 < did it ALL in COBOL. Not a line of C to be seen.   And that is progress?    <  < --  > < Paul Repacholi                               1 Crescent Rd.,9 < +61 (08) 9257-1001                           Kalamunda. B <                                              West Australia 6076, < comp.os.vms,- The Older, Grumpier Slashdot0 < Raw, Cooked or Well-done, it's all half baked.H < EPIC, The Architecture of the future, always has been, always will be. <  < --- ( < Incoming mail is certified Virus Free.< < Checked by AVG anti-virus system (http://www.grisoft.com).B < Version: 6.0.726 / Virus Database: 481 - Release Date: 7/22/2004 <  --- & Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).@ Version: 6.0.726 / Virus Database: 481 - Release Date: 7/22/2004   ------------------------------   Date: 6 Aug 2004 14:40:41 -0700 1 From: keithparris_NOSPAM@yahoo.com (Keith Parris) " Subject: Re: Solaris to Itanium...= Message-ID: <cf15391e.0408061340.44abe64a@posting.google.com>   d Paul Repacholi <prep@prep.synonet.com> wrote in message news:<877jsjsefg.fsf@k9.prep.synonet.com>..." > Read Bob Caldwells talk on that.  A My thanks to Keith Cayemberg for providing a pointer. Fascinating  lecture and  Q&A. Highly recommended.  & > The performance claims of the itanicE > it seems rested on an inner loop that was hand optimised for over a  > year.   8 Based on listening to the Bob Caldwell presentation, the? overly-optimistic performance claims you speak of, based on one B 30-instruction hand-written code loop, were made very early in theF project, apparently even before they had compilers (or at least matureC ones). That had to be way back in the '90s sometime. Today, we have B real hardware, real compilers, and real SPEC results. So that's no longer relevant.  B I found a lot more information in Bob's presentation that was very interesting.  D To give folks proper perspective: Bob Caldwell worked at Intel, mostF recently on x86 projects starting with the 2nd Pentium (P6) and ending> with the Pentium 4. He did NOT work on Itanium; all he had wasF impressions from a distance. He was one of 1 architects of the PentiumE Pro core. He admired Alpha. (Intel settled a lawsuit at great expense D over the Alpha intellectual property that went into the Pentium Pro,
 remember?)  A Up until the point he started working on Pentiums, he and several A others on the team had studiously avoided the x86, because it was > "ugly". There was no architecture model at all until P6 - just> reference implementations (and they often differed). So it wasD difficult to answer the question "What does it mean to be an x86?" AD C-code model developed for P6 because the gold standard inside Intel going forward.  C His complaints about Itanium were things like the large die size at C the time compared with Pentiums (yet Paul Demone shows that Itanium F core size today is comparable to all and even smaller than some of the: other chips out there like UltraSPARC IV or PA-8800 -- seeB http://realworldtech.com/page.cfm?ArticleID=RWT042704221446&p=5 ).  E And later in the talk, Bob points out that as transistor count grows, < one of the challenges is how to constructively use all thoseE transistors. One can only do so much with cache. Paul DeMone says "So E where is the state of the art in MPU design today with respect to the F hypothetical optimal performance CPU? Indications are that current CPU@ designs are still much simpler, smaller, and slower than optimalB performance CPU designs for their respective ISA and manufacturingK processes." http://realworldtech.com/page.cfm?ArticleID=RWT062004172947&p=5 0 So Itanium may just have been ahead of its time.  ? Bob was clear in the talk that preserving compatibility is very D costly, and that one can do much better starting with a clean slate.C He says orders of magnitude are possible if we used the transistors E for something more useful than cache (provided we can get the bits in ? and out -- and there I see the EV7-like integration as just the  ticket).  E It was clear from the presentation that Bob feels the x86 will die -- D that it is doomed to fail under its own weight due to complexity. InE the early days, each new generation of x86 implemented a single idea: A for the 486 it was pipelining; for the first Pentium it was 2-way D superscalar (two pipelines), plus branch prediction; for the PentiumE Pro it was out-of-order, speculative execution, etc. For later chips, @ there were no "single big ideas" left, so they had to do like 50A ideas, each gaining you 2%, and those tended to introduce massive C complexity and also interact in strange ways. You might fix one big @ performance problem and end up with two: one smaller problem andE another big problem you didn't discover until later. He compares this C complexity to a thoroughbred race horse - fast but fragile. He said D further speedups of x86 are "like pushing elephants up steep hills".  > He complained about Itanium using 1/3 of the die area (for x86F emulation) but having poor performance, and thought binary translation@ was a better path. Lo and behold, Microsoft's IA32 EL (ExecutionC Layer) takes exactly that path to better performance for legacy x86  code.   D He complained about power dissipation of Itanium, saying they had toF just lower the clock rate until they were under the power budget. (But8 process shrinks over time work in Itanium's favor here.)  A In the Q&A, discussing computer performance, he advises us to ask F ourselves, If we were starting from scratch, would we do it the way weD do now? and says: Throw out all the x86 stuff. Put a lot more effort	 into I/O.   2 That sounds to me a lot like Itanium with EV7 I/O.   ------------------------------  $ Date: Fri, 6 Aug 2004 19:27:53 -0400 From: "Ray" <no@spam.me>L Subject: Re: TECO manual location, was: Re: [Somewhat OT] What your non-OVMS0 Message-ID: <10h84vl1mor5vc7@corp.supernews.com>  8 Since you're using TECO, shouldn't this read "mung"? (^:   > --  E > I thought I would be the last on earth to mangle my e-mail address.  > fsnospam$elliott$$   ------------------------------  % Date: Sat, 07 Aug 2004 01:10:50 +0100 & From: Elliott Roper <nospam@yrl.co.uk>L Subject: Re: TECO manual location, was: Re: [Somewhat OT] What your non-OVMS1 Message-ID: <070820040110507456%nospam@yrl.co.uk>   H In article <10h84vl1mor5vc7@corp.supernews.com>, Ray <no@spam.me> wrote:  : > Since you're using TECO, shouldn't this read "mung"? (^:   You are right.   --  A I thought I would be the last on earth to mung my e-mail address.  fsnospam$elliott$$   ------------------------------  % Date: Sat, 07 Aug 2004 05:07:53 +0800 , From: Paul Repacholi <prep@prep.synonet.com>' Subject: Re: The free lunch YOU pay for 0 Message-ID: <877jscht5i.fsf@k9.prep.synonet.com>  5 Alan Frisbie <Usenet01REMOVE@Flying-Disk.com> writes:    > Paul Repacholi wrote:   F >> I would consider paying for the engineers free lunch the best value >> money of the lot.? >> Or as Ric said, "As long as your buying, I'll keep talking."    > s/lunch/beer/     :-)    That's the one!    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  % Date: Fri, 06 Aug 2004 21:18:31 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net>< Subject: Re: Time to patch OpenVMS - DCE-RPC buffer overflow+ Message-ID: <41143BF7.B8D71C7C@comcast.net>    Andrew Harrison wrote: >  > David J Dachtera wrote:  >  > > Undisclosed wrote: > > G > >>http://www4.itrc.hp.com/service/cki/docDisplay.do?docId=HPSBOV01056  > >>A > >>http://www.atstake.com/research/advisories/2004/a072204-1.txt  > >>B > >>patch here - http://www2.itrc.hp.com/service/patch/mainPage.do > >> > >>---  > >>= > >>hmm... seems that OVMS is NOT immune to buffer overflows.  > >>I > >>I'll be interested in reading the Nessus NASL script to check for the & > >>vuln when it comes out in 30 days. > >>? > >>damn DCE-RPC... that's the bane of Windows systems as well.  > >>9 > >>is the DCED installed as system or kernel on OpenVMS?  > >  > > @ > > ITRC is "login required"; so, I can't explore that just now. > > > > > Anyway VMS *IS* immune to *ALL* buffer-overflow attacks... > >  > 6 > That is not true its one of the OpenVMS urban myths.  E Really? I'm sure the folks at OpenVMS Engineering will find that most  interesting.  7 > It is relatively secure from buffer overflows but not 	 > immune.   & Hhmmm... interesting claim. read on...  ) > > ...until you startup a network stack,   H Praytell, Andrew: how does one launch a buffer overflow attack against a$ machine to which one cannot connect?  . Re-read the following until you understand it:& > > and then the vulnerability lies inC > > the network stack and/or related facilities/utilities, not VMS.    D.J.D.   ------------------------------  % Date: Sat, 07 Aug 2004 04:47:16 +0800 , From: Paul Repacholi <prep@prep.synonet.com>( Subject: Re: Touch Screen Voting systems0 Message-ID: <87fz70hu3v.fsf@k9.prep.synonet.com>  5 hammond@not@peek.ssr.hp.com (Charlie Hammond) writes:   3 > In article <87n01bc6r7.fsf@k9.prep.synonet.com>,  0 > Paul Repacholi <prep@prep.synonet.com> writes:  - >>What is the `problem' that is being solved?   = > The goal is to run elections that produce accurate, timely, ' > verifiable results with minimum cost.   ; And how will wince improve over a sheet of paper and a pen?    Costs are up a LARGE amount,/ The lawers will make `timley' seem a sick joke, " And you can't verify them to boot.E I suspect they will produce exactly the result that is wanted though.   D > The problem is that current systems do not do this.  As in Florida9 > 2000 and, although we hope not, possibly again in 2004.   A Some US systems do not do that. Large parts of the planet seem to 6 not have that problem. ( Got plenty of other ones... )  E > Although we won't have recounts in 2004 -- at least not in distrcts F > that use the Touch Screen system which cannot provide for recouts --? > the accuracy and verifiability of the counts are in question.   E How do you recount a computer poll? And how do you inspect the `Trade > Secret, comercial in confidence' crap lock inside out of view?   --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------   Date: 6 Aug 2004 11:01:52 -0700 $ From: paul.yee@heroix.com (Paul Yee)F Subject: Re: Ucx v4.0, network_mask set to 255.255.252.0 doesn't save.< Message-ID: <69011ae9.0408061001.5147b23@posting.google.com>  v "John Gemignani, Jr." <john@nfw-invalid.cibtrikker.com> wrote in message news:<aoednRZY_o9l647cRVn-rg@adelphia.com>...B > Did you just SET INTERFACE the address, or SET CONFIG INTERFACE? > $ > The latter will make it permanent. >  > -John  >   - Thanks, I just tried set config interface se0 F /network_mask=255.255.252.0 and "sho config interface se0" looks rightD but the running device is still 255.255.255.0.  So, I think this may7 be a problem with Ucx v4.1 and its ability to supernet.    ------------------------------  % Date: Fri, 06 Aug 2004 23:47:06 +0200  From: Dirk Munk <munk@home.nl>F Subject: Re: Ucx v4.0, network_mask set to 255.255.252.0 doesn't save.2 Message-ID: <cf0u91$dh5$1@news2.tilbu1.nb.home.nl>   Paul Yee wrote: x > "John Gemignani, Jr." <john@nfw-invalid.cibtrikker.com> wrote in message news:<aoednRZY_o9l647cRVn-rg@adelphia.com>... > B >>Did you just SET INTERFACE the address, or SET CONFIG INTERFACE? >>$ >>The latter will make it permanent. >> >>-John  >> >  > / > Thanks, I just tried set config interface se0 H > /network_mask=255.255.252.0 and "sho config interface se0" looks rightF > but the running device is still 255.255.255.0.  So, I think this may9 > be a problem with Ucx v4.1 and its ability to supernet.   O Yes, I think so too. A very long time ago I also tried to set up CIDR masks on  A UCX V4.x (don't remember the exact version), and it did not work.    ------------------------------  $ Date: Fri, 6 Aug 2004 17:43:06 -0400# From: "John Smith" <a@nonymous.com> E Subject: [OT]: $1 million prize for Ingres ISV's - How about VMS too? , Message-ID: <epidnXpDk-f0Zo7cRVn-rA@igs.net>  L http://story.news.yahoo.com/news?tmpl=story&cid=1738&ncid=738&e=9&u=/zd/2004 0805/tc_zd/132964   3 Open-Sourcing Rouses Interest in Cloudscape, Ingres    Thu Aug 5, 3:36 PM ET  Lisa Vaas - eWEEK   G IBM and Computer Associates made the best moves they could this week to I revitalize the use of relational database technology that would otherwise 3 just languish on the market, according to analysts.   B IBM disclosed at LinuxWorld in San Francisco that it is giving theC open-source software development community access to its Cloudscape K relational database by donating the code to the Apache Software Foundation. L This will allow open-source software developers to embed the Cloudscape code into their applications.  F On the heels of open-sourcing its Ingres database, Computer AssociatesK International Inc. announced a contest with a $1 million prize to encourage H open-source developers to create tools that will enable users to migrate? applications off of major enterprise databases and onto Ingres.   C Such moves likely represent the only way to keep these high-quality L databases from fading entirely from the enterprise software market, analysts said.   G Both IBM and CA have presented Cloudscape and Ingres to the open-source L community because neither of them were generating a lot of revenue from themI as commercial software products, said Noel Yuhanna, a senior analyst with & Forrester Research in Cambridge, Mass.  H CA's tools contest is basically a "marketing stunt" to remind developersC that "Ingres is another relational database that enterprises should L seriously look at," Yuhanna said. "Ingres has been a good database," but itsL use has faded in the enterprise market as a result of competition from IBM's: DB2, Oracle and Microsoft SQL Server relational databases.  H "These are products that have never taken off-products that don't have aL very good adoption rate," he said. "Getting these databases into open source= will perhaps trigger some greater adoption of these products.   G "Ingres has good features and functionality that give them an edge over F MySQL, which is now the dominant player in the open-source community,"I Yuhanna said. The contest gives open-source developers incentive to check  out these advantages, he said.       IBM reaps business benefits.  D IBM will benefit from the Cloudscape donation to the Apache SoftwareH Foundation by helping to support and expand the Java database community,F said Janet Perna, general manager of IBM Data Management Solutions, in Armonk, N.Y.  L The donation of more than 500,000 lines of Cloudscape code "would fill a gapH in the applications development area for an embedded database," which isD something the Java community has needed for a long time, Perna said.  I This in turn will likely provide business opportunities for WebSphere and L DB2 because the Cloudscape code is compatible with DB2, she said. This wouldJ allow Java developers to link Cloudscape applications to DB2 whenever theyD need to access an enterprise-scale relational database, Perna noted.  I Cloudscape, donated to Apache under its current product name, "Derby," is C particularly valuable to developers who need to build registries or 0 directories into their applications, Perna said.  J Click here to read about IBM's decision to release the Cloudscape database to the open-source community.   D Cloudscape also will help provide new incentives for Java enterpriseL applications development, said Ken Coar, an IBM senior software engineer and- an Apache Software Foundation vice president.   H "Java is not considered to be the choice in terms of performance" in theG database area," Coar said. "Cloudscape will give the lie to that image" J because it is a "very efficient and lightweight environment" that is built as a set of Java classes.   H It has all of the features expected in an enterprise database, includingI support for the SQL 92 standard with support for transaction triggers and L two-phase commit functionality, which are essential for reliable transactionE processing, Coar said. Cloudscape essentially provides "high-end data , processing brought to the low end," he said.  D With Cloudscape, Apache and the open-source community received theseK blessings: the biggest chunk of code released so far; the first relational, I embeddable Java open-source database; and reason for a lot of open-source  developers to be very happy.  G "Are we excited? Yes," said Greg Stein, chairman of the Apache Software G Foundation. "[Cloudscape] makes it very easy to deploy certain types of C applications in the low- to midrange type of database, which is not > satisfied by anything else in the open-source community today.  L "By donating that code to Apache, a lot of people needing that functionalityJ can now get it, and under the Apache license, which is a very nice licenseK to work with, any user-[be they] hobbyist or commercial entity-can use it."   E Stein said the open-source community has lacked a fully featured Java J database and instead has relied on connections to MySQL, IBM DB2 or Oracle Corp. databases.  A "What's interesting in this one is it's embeddable into your Java L application," said Stein, in Palo Alto, Calif. "You don't have to connect to a separate server."       ) Streamlining Java application deployment.   D That will make it easier to deploy applications based on a Java codeK package, Stein said. "It makes it a lot easier to build and deliver systems J based on this code, compared to a classic database like Oracle or DB2," he said.   F In particular, the fact that Cloudscape is an embedded relational JavaK database makes it superior to comparable embedded databases such as the one J from Sleepycat Software Inc., Stein said. "This one really brings the fullJ functionality of a relational database embeddable into Java applications," he said.  L Stein foresees Cloudscape being used in small to midrange applications whereJ ease of deployment is of concern. It will be usable both on clients and onB servers, since it's embeddable and can be incorporated into client
 applications.   F "There's certainly going to be a lot of hobbyist activity, [plus] someG commercial activity around people wanting to incorporate this into some # lightweight applications," he said.   F The code will be sequestered in the Apache incubator project until theJ foundation determines that it has all property rights to the code and thatG it has been granted the right license to use it, to relicense it and to  further develop it.   I After that, the incubator's role is to build the community around the new D code base. That's to ensure that the project outlives any individualJ contributor, whether it's an individual developer or a group of developers& assigned by a corporation such as IBM.  J "One thing about Apache Software Foundation is we want to have communitiesF set up around each of our projects," Stein said. "We want to make sureC there's a large group of people actively involved with the project, K supporting it and continuing work, so the project has long-term viability."   K IBM pitched Cloudscape's release as an altruistic move, and its high regard L in the open-source community helps bolster that pitch, regardless of how IBMD might benefit from disrupted market. Specifically, analysts point toK potential victims as being Microsoft Corp. with its technology stack and/or + Sun Microsystems Inc. in the realm of Java.   L "IBM has been quite good about ensuring that they work with the community asH opposed to trying to manage it or take over existing communities," SteinJ said. "They're very much just another developer in the community, which isG something we like to see at Apache: a meeting of equals as opposed to a B meeting of leaders and underlings. We like everybody to be equal."      J Check out eWEEK.com's Database Center at http://database.eweek.com for the+ latest database news, reviews and analysis.    ------------------------------   End of INFO-VAX 2004.434 ************************                                                                                                                      