Anyone read the latest online DB2 Mag's articles?

Has anyone read the online DB2 Mag's article on the Informix benchmark 'debate'? They've published Kate's survey... I saw that there has been a bit of debate on IBM performing an IDS benchmark, both in Eric's blog and Kate's survey. Since I am one of those survey responses (only 4 votes!? :-( :-) ) and I have been one of the one's screaming loudly for this, I wanted to address this issue and explain why I wanted benchmarks and with the response from Jerry Keesee on the last Talk with the Labs conference call, why benchmarks may not be the best thing. First, I think its great that there has been a lot of response to the IIUG survey regarding benchmarks. Sure its true. No one based their purchasing decision on a standardized benchmark. Ever since Oracle got caught rigging their software on the early rounds of TPC benchmarks back in the 90's, people have assumed that all companies are doing things to skew the results in their favor. So while its not something that will make or break a company's decision, it is something that is important because it get's your product on the radar as a viable choice. In an online developer's forum survey, Informix was barely on the radar, while DB2 didn't make the cut. The problem facing IBM is that unless they can make their database products relevant, they are not going to be asked to participate in the sales cycle. IBM sales teams can and will sell Oracle in place of DB2/IDS if it means an easier sale of other software and of course IGS/BCS services. On the core S&D team, there is no incentive to promote IM Database products. Out of sight, out of mind. All it takes is one Oracle bigot or a bias to Oracle and the S&D team sells Oracle. I think that's the major issue. Getting Informix on the radar of the consumer as a viable option. Its the first step in getting the product sold. The purpose of the benchmark is to open the debate. To show that on comperable hardware, you can expect better results with Informix. Of course by comperable, there are only two numbers that make sense. The volume of transactions and the cost per transaction. So if a customer needs a system that can handle a sustained 10,000 tpm, they can expect that the solution would cost X. 100,000 tmp, it would cost Y. The fact that the benchmarks are audited (or so they should be) and independent of the vendor, the consumer can put more faith in these numbers. And of course, when the benchmark is released, its a PR piece that can generate more news and again the more news, the greater the chance to get on the public's radar. The benchmark is really chest thumping at it comes at a price. The reason that I've questioned IBM's benchmarking practices is that it is not in alignment with their stated marketing message. Of course since pointing this out, IBM has realigned their marketing message, and Jerry Keesee's response to the question I asked on the call has answered some of the issues. First, it looks like IBM has started to walk away from the 'IDS is for OLTP and DB2 is for DW'. The current product announcements and future road map show IDS adding features that would make it viable as a DW and would become more competitive on this front. So Jerry's presentation shows IDS and DB2 as an either or proposition. Meaning that both databases can be used in either situation equally as well and its up to the consumer to choose. Jerry also made the point that IBM has already invested heavily in to DB2's benchmarks. They have an existing team and already have a series of numbers over time. Informix hadn't done any benchmarking in the past 10 years so that there would have to be a net new investment in to doing IDS benchmarks. Then there's the issue of which benchmarks. Since IBM is committed to doing DB2 TPC-C benchmarks, it doesn't make sense to release IDS benchmarks that beat DB2. Jerry also pointed out that on the TPC-E benchmarks, there is only one database. SQL SERVER from Microsoft. Without a competitor, you don't have any benchmark wars and you only have a 'hardware war'. A point Serge made earlier is that the benchmarks are for the hardware companies sake and not the database. There is some merit to this argument, and by avoiding a pitched war with Microsoft, IBM avoids committing a revenue stream on these benchmarks. Jerry also made a point. IBM was reluctant and more to the point, unwilling to commit the resources for an IDS benchmark and the additional revenue commitment going forward. One thing I learned from working within the blue pig is that you need to pick and choose the battles you are going to fight. If IBM isn't willing to make the committment then you're not going to be able to change their minds. (Unless you happen to be a large hedge fund shareholder, or a Wal*Mart client, neither of which have a vested interest.) While answering my question, Jerry opened up a small can of worms. Just how much does it cost to do a benchmark? I mean the IBM hardware isn't cheap, but its internal and there isn't a real cost to IBM to access this hardware. Same too for the software. So what's the real cost? Labor? Yeah, dedicated benchmarking teams don't come cheap and with all of the cost cutting, even moving the team offshore isn't going to help. (These have to be your best and brightest... ;-) So suppose the money to do a benchmark is $X (USD) and the annual cost going forward is some f(X). How much of a commitment is IBM willing to make in IDS marketing? After all, that's the real purpose of the benchmark. Getting the product on the customer's radar. Now both DB2 and IDS are brands. So if IBM is only willing to market brands, how much are they willing to commit up front? How much in a continuing basis? I think those are the relevant questions that you need to ask IBM. Focus less on the benchmark, but on how much revenue is IBM willing to commit to promote the Informix brand? (You'll get that chance next month if you attend the IIUG event in KC.) And now we're getting to the real kicker. If you had that money, what would you do to get Informix out in front of the consumer? Maybe that's the next question for Kate and the IIUG to ask? After all, the IIUG *is* IBM's best marketing arm for IDS... ;-) But hey! What do I know? I'm just the guy who's been screaming the loudest and longest on this issue. I just wanted to take the time to explain why a shift in IBM's positioning and response makes it important for us to shift our focus in a more positive direction and to avoid a confrontation. -G

On Mar 12, 12:33 pm, Ian Michael Gumby <im_gu...@hotmail.com> wrote: *is* IBM's best marketing arm for IDS... ;-) > > But hey! What do I know? I'm just the guy who's been screaming the > loudest and longest on this issue. I just wanted to take the time to > explain why a shift in IBM's positioning and response makes it > important for us to shift our focus in a more positive direction and > to avoid a confrontation. > > -G I wanted to follow up on Eric's blog too. I agree with Eric that while a benchmark isn't the deciding factor, over 1/4th of those surveyed indicated that they do pay attention to benchmarks during their decision making process. I take this to mean that having a benchmark is going to put the product on the consumer's radar. Customers tend to prefer products that they've heard about in the news. Very few customers are going to go out on a limb and purchase a product where they can't easily google and find a lot of information on the product. With respect to the TPC-E benchmark, at one point IBM 'promised' the benchmark. However Jerry K's latest statement is that IBM doesn't want to get into a benchmark war and with only Microsoft as the only player, you can't get in to a war. I know I'm reiterating a point made above, but that's a very important point. IBM is thinking long term and they don't want to spend/waste money on something that they can't show a direct correlation between the monies spent and the increase in sales. While we all agree that something needs to occur that will place IDS on the consumer's radar, we don't agree on what it will take and how much of an investment should be made. Of course that's really IBM's choice because its their money. We as the community can only voice our displeasure or opinions when IBM fails to take any action. (While inaction is a valid action, in this specific case, inaction is an indication of a lack of confidence in IDS as a viable product moving forward.) The bottom line *is* product awareness. Without it, all of the investment made by IBM in to IDS and their future road map is a wasted effort of talent. -G PS: Note to IBM.... Even grass root efforts cost money.