Friday, December 24, 2010

Some blog postings for year end

We don't really celebrate Christmas here, but it is the end of the year, and we need to finish our vacation days, and since our colleagues from overseas are now in vacation,  it is a good time to do it, so it is a good time to do catch up on things that are not part of the regular work.   Looking at some recent Blogs,  I borrowed the above picture from Scot Fingerhut's Blog entitled:  " How Santa Uses CEP - Elf Productivity, Real-time Naughty And Nice Rating" .    BTW - I found another Christmas related postings in OMG facts claiming that Jesus was not really borne in December,  

Back to the professional Blogs, my IBM Haifa Research Lab colleague Yishai Feldman started a Blog entitled "The Dusty Deck"  with punched card on the top -  I belong to the generation who really punched cards, I guess that today punched card belongs to museums along with typewriter and sliding rule. 
Yishai is a local software engineering guru,  who deals with analysis of programs, so I'll follow his Blog with interest.

More in the Blogland,;

The Prova Blog reports on a new version of the Prova open source event processing system that was released this week. 

Colin Clark brings us predictions about 2011.   About event processing he predicts that it will become part of a larger horizontal offering.  I think it is already happening,  as I stated before there is still a niche for the "stand-alone" event processing,  the major market is in having event processing capabilities pervasive over computing offerings.

Last but not least, Rainer von Ammon summarized the workshop he organized recently in conjunction with ServiceWare 2010.   





Sunday, December 19, 2010

S4 from Yahoo Labs


Yahoo Labs recently announced an open source called  S4:  Simple Scalable Streaming System, available as open source.  A paper describing S4 is available, as well as a site from which one can download the software (alpha version). 
This is a distributed platform whose model is a kind of event processing network; the stream dataflow model where each node is  "processing element", similar to IBM Infosphere Streams;  The paper states that it was inspired by the MapReduce model, but I still need to read the details.  It is interesting to follow the branch of distributed event processing platforms.  It seems that this is a "programming in the large" model, where each processing element is written in Java (at least according to the paper's examples) it does not have its native language.