Re: Help - Does DSMS/Data stream fall under "DBMS design and implementation" category?

From: Marshall <marshall.spight_at_gmail.com>
Date: 4 Oct 2006 09:22:19 -0700
Message-ID: <1159978938.947821.260490_at_b28g2000cwb.googlegroups.com>


Jan Hidders wrote:
> Marshall wrote:
> >
> > I agree that this is simple in principle. However, if I am
> > understanding
> > correctly, this would seem to put some severe constraints on
> > implementation and thereby scalability, which is one of the points
> > of a DSMS.
>
> Well, that's why it's called research, Marshall, they try to make
> possible what at first sight seems not. :-) In many cases you can use
> certain knowledge about the stream like the fact that it may be ordered
> in some way, or nearly ordered within some bounds (think of packets
> that almost arrive in the order of their time-stamp, but with an error
> within a certain bound).

In other words, if one can bound the join attributes on both streams, one can join on those attributes and only have to store the stream within the bounds. Interesting. I guess that's why all the examples were joining on timestamp columns.

Well, anyway, I should read some of their papers.

Marshall Received on Wed Oct 04 2006 - 18:22:19 CEST

Original text of this message