To: Punko who wrote (2026 ) 6/22/1998 4:56:00 AM From: GUSTAVE JAEGER Read Replies (1) | Respond to of 3194
I agree with you: comparing the growth rate of a multibillion $ industry with a niche's that's only worth $100M doesn't set a good ''benchmark''... Yet, you have to take into account the economic seachange that's swept through the whole IT industry: who'd dare to imagine today that another MSFT is poised to emerge from scratch? Or another ORCL or even another Computer Associates? The so-called entrance ticket to the ''XXL size'' team is much more expensive today than it was 20 years ago! Just ask Netscape!! Now, focusing on the DBMS market, we can see that the intrisic inertia of this market seems much more important today than it was 15 years ago: what was the total investment dedicated to DBMSs by US corporations 20 years ago compared to what their current DBMS expenditures represent today? Hopefully for ODIS, it seems that this corporate inertia is losing steam: hybrid ORDBMSs are no longer a serious threat to pure ODBMSs (see the article below) and even if ODIS, along with other ODBMS vendors, doesn't show a 60+% growth rate yet , the fact that established RDBMS vendors are undergoing a sluggish growth in the while is very encouraging. For if both RDBMSs and ODBMSs are still enjoying a 40+% growth rate, where's the paradigm shift??? As I said it earlier, there's a twofold scenario going on: first we need a shrinking RDBMS market, followed by an accelerated ramp-up in the ODBMS market. Here's an interesting article that I found on techweb.com . What's funny with this article though is that its suggested message is: "Don't try the ORDBMS way, it's not mature enough and it's too expensive with regard to its potential ROI... and don't try the pure ODBMS avenue because that's really pure folly! Bottom line: just stick to your good old RDBMS..." June 22, 1998, TechWeb NewsA Trickle Of Objects -- Interest in object-relational databases has waned sharply By Rich Levin Remember when the smart money was on object-relational databases? Just over a year ago, vendors and pundits alike were touting the universal database, and promises of flexible application systems that adapted rapidly as business needs evolved were plentiful. Fast forward to summer, 1998. Instead of a tidal wave, object-relational is more of a trickle, dribbling into IT through a leaky faucet. The universal database --eagerly touted as a key enterprise system that would let organizations mix and match conventional text and numeric information with business objects, components, and varied multimedia data types-- hasn't lived up to early expectations. And while overall adoption of object-oriented technologies continues to rise, adoption rates for object-relational databases are downright anemic. Universal databases were positioned as more flexible than so-called "pure" object-oriented databases --which have been around for years-- because the universal databases extend existing relational technology, making it easy to port legacy online transaction processing (OLTP) systems to new object-relational platforms. The idea was that strategic data would be accessible to new and existing applications developed for the Web, E-business, extended supply chain, transactions, and multimedia requirements. But few companies have bought into the notion that object-relational databases are the fastest way to make operational information available to analysts and trading partners. "Object-relational is not a high priority in IT today," says Jim Aviles, senior VP for merchant services with BankAmerica. "Embracing technology initiatives used to be seen as a value-add, even when they didn't have demonstrable bottom-line impact. Today, more companies pass on new technologies until they can clearly understand how they're going to return economic profit. That aspect still isn't clear with universal servers." A survey conducted by the International Oracle Users Group America in February found that nearly half of respondents are using or plan to use distributed object database architectures, yet only 25% were "very interested" in storing those objects inside the universal database. Conversely, 70% of respondents were not interested or only "somewhat" interested in deploying universal database applications. Weak interest in object-relational databases isn't limited to Oracle users-Informix and IBM, early trendsetters who introduced universal database strategies that promised to jigger traditional wares for object-relational use, have had trouble selling the systems for multimedia applications. "It's an upgrade strategy the database vendors glommed on to, and it ended up being a total failure," says Michael Barnes, an analyst with the Hurwitz Group. "The people who need the functionality to support multiple object and multimedia types are using pure object databases, not the universal hybrids. And that's turned out to be a small market." Wrong Place, Wrong Time Bad timing is another reason that object-relational excitement has slumped. "People on the database administrator side have been overwhelmed with year 2000 and haven't had the time to invest in learning object-relational," says Rich Niemiec, executive VP of the Oracle users group. "CIOs are also distracted. They're scared to death not only of the year 2000, but that their CEOs don't understand the risks of not taking on the Internet. They are wondering how they can stop the next Amazon.com from cannibalizing their business." IT leaders agree. Many say they've been frustrated by the failure on the part of universal database vendors to deliver a sound economic model to justify relational system migrations. No one doubts the tech- nology's advantages, but few have been able to sell its return on investment. As with previous overhyped technologies, object-relational is a technical success and a market failure. [...]