Sybase data integration

Sybase is not normally thought of as a data integration company. True, it has offered two-way heterogeneous replication (which does not need to include a Sybase source or target) for some time but with respect to other aspects of data integration it has not been a significant player, despite its acquisition of Avaki (a vendor in the data federation space) last year. Nevertheless, last month Sybase released its data integration strategy.

The Sybase Data Integration Suite builds on the two capabilities mentioned in the previous paragraph and adds real-time event capability through log-based change data capture, search (through last year’s acquisition of OmniQ), a common administrative environment and modelling (via PowerDesigner) and integrated metadata management through Sybase Workspace. All this is in the first release and it will be followed in the first half of 2007 with version 1.1, at which time ETL (based on the recent acquisition of Solonde) will be integrated within the suite along with a number of other new features and capabilities.

Now, from a traditional data integration perspective there are a number of things that are interesting about this; in particular that Sybase is including search and modelling within its environment but has yet to announce either a partnership or acquisition for data quality provision. It is also notable that Sybase calls its administrative tool the Sybase Data Services Administrator.

Herein lies the key to what Sybase is doing: yes, it is building a data integration suite with the sort of range of facilities that one might expect in such an environment (data quality, in one form or another, can only be question of time) but it is also building a direct competitor to IBM’s Information Server—while IBM refers to information services, Sybase prefers data services—same difference. The only real question is whether Sybase will add master data too: it wouldn’t surprise me.

So, how good is the Sybase Data Integration Suite when compared to Information Server? And how much impact will it have on the market?

Compared to IBM, Sybase has ETL and federation products that are much less mature more than those of IBM’s and, while I don’t claim to be an expert on search, I would guess that that the same applies to OmniFind versus OmniQ. Also, the Data Integration Suite doesn’t haven’t a common metadata repository yet (it will have). On the other hand PowerDesigner has much more capability than Rational Data Architect and I would expect Sybase Replication Server (which has more than 2,000 customers) to be able to give IBM a run for its money at the very least: so, I guess any comparison will depend on where your priorities lie.

Perhaps a more interesting question than competitive comparisons between IBM and Sybase will be the effect on the market. How, for example, will this impact on Informatica? Can we expect to see modelling and search incorporated into its offering? My guess is that it depends: I usually work on the basis that one new thing is interesting, two is a co-incidence and three is a trend. So far we are at the co-incidence stage but it will only take one more vendor to move into the information/data services arena and we will clearly have such a trend. As I expect that this is precisely what Oracle intends to do with its acquisition of Sunopsis then this is the direction I expect the market to move in.

What is most interesting is that it is Sybase that has joined IBM as first movers in this market: you could be forgiven for finding that a surprise, but good for them!