An interesting meeting with IBM

Folks, Yesterday we had an interesting techo/positioning meeting with some IBM staff. The attendees were nearly all locals (Australia) excepting a representative of the IDS development team from US. The IBM'ers were a mix of techies, managers and supply chain staff; from both sides of the aquisition (ie some were long-term Informix, some long-term IBMers) These IBM Informix presentations have been targeted at the top 100 Informix partners globally. 10 of these were based in the Asia/Pacific region. I think meeting with 100 partners was just an arbitrary limit of practicality. Hopefully this posting will share a lot of details with the smaller players. I'd also like to hear any alternative perspectives from anyone else who has sat though one of these meetings. The meeting predominately discussed IDS and DB/2, WebSphere and 4GL, and a small admission of slackness with Informix positioning (!) and recent supply issues. I don't believe there are any trade secrets in here, although I suppose some points could be exploited by rabid members of other DB suppliers. Let he who is without sin, cast the first stone; for example, before someone from Oracle can gloat about the words here regarding ease of admin of DB/2, they need to consider the cost of administering Oracle... This posting is largely based on our internal note-taking which has been passed around. I'm posting it in the hope that it will help to strengthen the Informix community. ** Engine, General info - IBM welcomes working with partners with regards to feature requests, and the possibility of implementing these into future releases. - Defect Backlog has dropped to about 15% of the count just prior to the takeover [15% is my guess from the graphs and numbers tossed into the air]. Note that the peak of defects around 2002 was partly due to the very major work that was required to create the Universal Server features (UDT, datablades, etc) - All customers are very strongly encouraged to upgrade to the latest versions due to the major reduction in the Defect Backlog - IDS development staff have been given a large increase of resources. - IBM afforded the purchase of a very large testing machine farm, so that testing now applies 60,000 tests on 9 different platforms within a week; a task that used to take 3 months (only on Solaris) prior to IBM. These 9 platforms are now teir 1 platforms, as opposed to the older model: Solaris, then 5 teir-one's, then the strays.... - The tests are applied across enough machines to ensure that virtually every test is applied overnight. The tests are striped around the machines so that each machine applies all tests within the week. - With the nightly tests, if a new piece of code is checked in that affects performance by as little as 1% in some area, it will be automatically flagged for follow-up the next day. - IBM policy demands that no release is made if there are any level 1 and level 2 severity defects. Sign-off will not be granted unless this can be proved. - IDS has a development plan that takes it to version 9.8 in 2008; beyond that, nobody plans any further ahead... - Development of new features in IDS has accelerated (approx 40%?) since IBM - Uptake of the 9.40 engine has been the fastest ever, for any upgrade ever released - Approx 50% of all 9.40 deliveries has been for Linux. - IDS has received a lot of code from DB/2. For example, in 9.4, the GLS system was revamped with UNICODE source taken directly from DB/2 - Multi-level security will be brought in from DB/2. - IDS will be progressively enhanced so that it integrates very tightly with the entire IBM product line; eg Tivoli, Websphere and other products I'm only dimly aware of. "Software Group Integration" ???. If another product applies SQL's that are not handled vvery well with IDS, then either IDS or the product or both will be optimised. - IDS is giving a lot of source code to DB/2. For example, HDR code is being added to DB/2 instead of re-inventing it. - Both engine teams are sharing optimiser ideas. - IBM is striving to emulate the simplicity and ease of administration that IDS has. By 2008 they plan to make DB/2 as easy to manage as IDS. ** Planned IDS 9.50 Features - Further improvements to "Safety" and HDR (ease of use in particular) - Further improvements to increase availability (ie during admin) - Autonomic admin (eg self-repair of tables with excessive extent fragmentation, automatic update stats) - Integration into IBM toolset - Multilevel Security (sourced from DB/2 as recently released in v.8) - Faster restarts - Table level Point-in-time restores - Query/SQL plan tracing tools - Backup/Restore (??? what does this mean?) - Native .NET driver - MS Visual Studio integration - Java 3.0 (JDBC) compliance. This requires quite a number of added features just to be able to achieve full JDBC compliance. - XML Shredding / Publishing. XML will be storable efficiently, and Xsearches will be optimised and efficient. - Web portal / sphere integration (certified release) - SQL standards compliance - ESQL/C to DRDA integration (i.e. re DB2) - DRDA integration within IDS server (fully integrated by IDS 9.6) - Transaction / 2-Phase commits for MQ series - Better Tivoli integration, possibly with a bundle of "Tivoli Lite" to replace ISM - Encryption - down to column level. - Variable PAGE size upon initialization - 2k, 4k, 8k, 16k or 32k. - Default roles for users - Single user admin mode which allows access to data via SQL for privileged user. - Rename dbspace (this is in 9.4.UC3) - Better role separation re DB admin - Truncate table (Jonathan, did we win after-all on the logging issue?) - Optimizer improvements query plans (i.e. 9.40 was optimized more for OLTP) - Greater SHMEM - Extensibility in UDR's, including multi-engine use - Refined installation, allowing to only install components of the server, i.e. reducing foot-print of install. - Improved admin tools with more alerts and alarms ** Planned IDS 9.60 features (2007) - Any customer/partner requests - Improved "safety" - DB/2 interoperability - engine to engine and more shared syntax and features - Complete certification of integration with IBM product line - Better embeddability - ie ability to inject a tailored install of IDS into an Application install, effectively allowing developers to shrink-wrap the engine with the application. ** Planned IDS 9.70 features (2008) - Any customer/partner requests - Getting vaguer now... - At this time, they hope that DB/2 will be cabable to be installed as an upgrade of IDS in a process that is as simple as upgrading IDS versions. ** IDS after that - The forecast is now too far into the future to make any sensible plans. ** WebSphere - v5.1.2 includes a 95% port of 4GL syntax into WebSphere. - v5.1.2 is not considered quite complete enough to go into produ

Thanks Andrew, very interesting. > - All customers are very strongly encouraged to upgrade to the latest > versions due to the major reduction in the Defect Backlog Is it just me, or is there a flaw in IBM's logic here? If IBM wishes to "strongly encourage" users to upgrade to v9.x (and I apologise profusely for the necessary split infinitive here) wouldn't this aim be best served by scrapping the upgrade charge? I still maintain that 90% of users have no need of the additional functionality of v9 and will therefore hang on to their cash until forced to upgrade, at which time much resentment and consideration of alternatives will ensue.

Captain Pedantic wrote: > Is it just me, or is there a flaw in IBM's logic here? If IBM wishes > to "strongly encourage" users to upgrade to v9.x (and I apologise > profusely for the necessary split infinitive here) wouldn't this aim > be best served by scrapping the upgrade charge? I still maintain > that 90% of users have no need of the additional functionality of v9 > and will therefore hang on to their cash until forced to upgrade, at > which time much resentment and consideration of alternatives will > ensue. OK - i didn't mention the reason for the urgement (have fun with that one) but it was simply that the defect count is so much lower. Getting the latest 7.31 (.UD8 about to be released, iirc) was also suggested for the same reason. As for the pricing, since 7.31 is no longer undergoing new feature development, I guess pricing needs to be dealt with as you suggest...

Great post Andrew, a true service to the Informix community? Just a couple of bits of shrapnel to throw around: - Thank goodness they seem to have finally done their homework on IDS. Let's hope the apparent uprating of its perceived value will be translated into some half-decent marketing and customer service. But there's still a tacit assumption that everyone will one day go to DB2. Everyone knows that simply hasn't been happening (the site I'm on at the moment is gradually moving to Oracle and my previous one is in the middle of a wholesale switch to MSSQL). They've got their heads where the sun don't shine if they think this drift will be changed easily. - They might as well rename the damn thing. Then they'd be able to keep it as a separate product without shuddering at the thought of "that" takeover. IDS is a cr*p name. OnLine was better. Suggestions .... ? - As one for whom the Informix product line has provided 90% of my ability to feed myself since 1986, my main personal concern is that my skills are still going to be worth something after 2008. Even if IBM managed to rein in their remaining sceptical users I'm still not convincved I'll have a job / client base if everything's going to have DB2 written in front of it in four years. I don't know what they can do about that but I hope they will recognise that Informix has a loyal, highly-skilled base of technicians. I hope that won't all go in the bin. Frankly if I *DO* have to re-skill and start back at the bottom rung I'd still be inclined to do so in Oracle or MSSQL. Otherwise - sheep farming anyone? Andy Kent

"Andrew Hamm" <ahamm@mail.com> wrote:- > Let's hope so. I think the biggest problem for Informix developers is simply > the false perception that Informix is a risky investment. As many people > have pointed out, prospects are very unlikely to read a white paper buried > on a website, even if they were motivated to go looking for one. There is a strong reason for this false perception. In 3 yrs since IBM took over Informix, there has been no marketing of Informix. Can you point out a single news source of new informix customer wins after IBM took over. Either it means that not a single new win has occurred or IBM just doesn't care. Now compare it with DB2 promotion. Interesting meeting like yours means jachshit to Informix customers. It has no value.

"Andy Kent" <andykent.bristol1095@virgin.net> wrote in message news:ac5fb36.0404070543.66cbd079@posting.google.com... > - As one for whom the Informix product line has provided 90% of my > ability to feed myself since 1986 ... I hope they [IBM] will recognise that Informix has a > loyal, highly-skilled base of technicians. I hope that won't all go in > the bin. I think you can rest assured that the interests of Informix technicians will be of paramount concern to IBM.