I’ve spent the last couple of days, after absorbing the Sybase acquisition
of Aleri’s assets, looking at the various CEP players websites for evidence
of cloud (grid name du jour) deployment.
I haven’t found anything worth mentioning. That’s not to say that I
might have missed something. But my guess is that with all the economic
woes lately, that the CEP vendors have either chosen to ignore or just simply
aren’t aware that the rest of the world is actually solving real problems
with cloud deployments.
Which is interesting.
It’s interesting because one of the pioneers in CEP & database space, Mike
Stonebreaker (love that name), has gone on to work in stealth mode (it says
stealth on their web site anyway) over at VoltDB. VoltDB is based upon work
done on HStore and would appear, on the surface anyway, to incorporate all of
the things we’d expect in an OTP environ... (more)
VMWare has been on a buying spree lately.
In the last month, they’ve announced both Redis and RabbitMQ.
Here’s VMware’s take on Redis, and spring source’s on RabbitMQ.
RabbitMQ is built with Erlang.
Much Rejoicing in the Village
We use both of these technologies at Cloud Event Processing. And we love
Erlang too. VMware’s acquisition of these technologies not only validates
our decisions, which we are very selfishly pleased about, but also sends an
The Message Please
Everyone’s busy abstracting resources in the cloud – making resources
like compute, stora... (more)
CEP isn’t really about low latency. The ability to do things quickly is
important, just as in any system – especially those systems that grow and
need to handle a lot of information. Doing things quickly means doing
things efficiently. And doing things efficiently means less money spent on
hardware. Theoretically anyway.
SO WHAT IS REALLY COOL ABOUT CEP?
CEP gives one the to submit queries like “select symbol, avg(shares) from
trade_stream group by symbol over 5 minutes emit every 1 minute.” The CEP
engine would consume this query, and then start returning an average of
Well, it would appear that Michael Stonebraker may have hung up his research
hat and joined the marketing team.
First, read this, “Will the Real Column Stores Please Stand Up?”
And now read my reply, which Vertica has yet to approve on their blog.
Mike, You’re describing an implementation, not an algorithm. An analogy
would be saying that anything that didn’t look like Hadoop wasn’t
map/reduce. Column stores exist because disk is slow – not because
they’re some new and magical way to store data. As a researcher, you’re
well aware of the fact that it is impossible to prove that a... (more)
On this continuing series, I am examining thoughts and specific
implementation details around building a back-testing platform for algo
trading. Eventually, we’ll see where complex event processing plays and
how to implement it.
Appendix to Part One – The Data Format
Rather than looking at various database solutions first and then trying to
define the problem in terms of those solutions, let’s first examine what
market data looks like. In its most simple form, market data looks like
this (there’s usually a little more, but this is fine for our purposes):
Date: The date of the ... (more)