To: nanocap who wrote (445 ) 11/29/1999 11:20:00 AM From: Daniel Chisholm Respond to of 477
Hi nanocap, Regarding the technology issues,...scalability is an advantage with VSNT. Look at applications and transaction processing requirements of latest clients (e.g. France Telecom, Ericsson, GE/Harris, etc.). These all require very robust, reliable and scalable ODBMSes. Two points on this. First, I don't think that "scalability" really, actually matters for *most* applications -- by this I mean that *most* applications will inherently be "small" or "mid" sized. A company can have a product that "breaks down" for terabyte-sized databases, yet be a perfectly usable and sound product for 95% of the applications out there. The only real downside I can see to having a "not scalable" solution is that you can't succeed at being the "OODB of choice" for the big, sexy applications out there -- so you do give up a bit in the way of bragging rights. On the other hand, the big, sexy "glamor" applications do not necessarily translate into big revenue for an OODB vendor (sometimes bragging rights are expensive!). If Versant was generally agreed to be "not scalable", then I (a skeptical short seller!) would *not* hold it against them -- provided they recognized this, and just went out and made a fortune selling their OODB for bread-and-butter type applications. The second point is that even though the client list you mention are big companies with many big and fancy requirements (e.g. reliability, real time performance, etc), that does not necessarily mean that that is what they are using Versant for (parenthetically, I thought that most OODB vendors were desperately trying to escape the "Finance/Telecom Ghetto", and break out into simpler, bigger, more profitable applications?). I know of several "sexy" applications for OODBs (Australian telecom 1-800 database, NY Stock Exchange data) that sound impressive, and are certainly worthwhile achievements, but they don't really count for as much of a "super stress" scenario that the OODB salesmen might have you believe. There are many seemingly mundane applications or use cases that can bring the best OODBs to their knees.object locking vs. page locking. This is a very much discussed architectural point that each side (Versant: object locking, ODIS: page locking) likes to blather on about continually. There are advantages and disadvantages to each, of course. When all is said and done, it is my opinion that both approaches are OK, and that neither criterion is sufficient in and of itself to cause one solution or the other to be "inadequate". The OODB salesmen will surely beg to differ of course (that is their job ;-), but I don't think this should cause anyone to dismiss either product out of hand. It is my assessment that an application developer can build a good product with any of a number of the OODBMSs out there (if you can't, it probably points up a weakness in your design!). FWIW. - Daniel