Israeli Hi-Tech tsunami is here?

Just read about Qlusters R.I.P. They left this world with 11M$ in the bank. The company just recently closed a 10M$ round.
So, what's the story? Well, Qlusters developed a very promising technology and gained some awards for their achievements throughout the years they existed.
The company raised quite a lot of money and its burn rate was not so fast, at least as I can estimate according to public numbers.
And still, the company investors decided to take back their investment and shut the company doors. This means not only the investors couldn't see how the company works out the near-coming-bubble-blow, but they stopped believing in this company ability to exist in the market. Just after they decided to invest...

Well, always look at the bright side of the moon....

ESB or BPEL?

Yet another interesting question I tend to get asked about almost in every SOA panel. There are many flows that can be implemented as mediations (ESB) as well as BPEL (BPM).
Almost all vendors bundle ESB and BPEL engines in same product edition which allows even better flexibility to the architect / designer to decide which technology to use for the implementation.
Examining the SOA architecture, it is quite clear that BPM and ESB are there for very different needs - BPM for the business processes requiring analysts to model and business owners to monitor, while ESB for integration specialists to mitigate protocols, models and technologies.
Marc Fasbinder, BPM Integration Solution Architect, IBM, gave a nice summary of features and differences between BPEL and ESB which gives some key directions in choosing the correct technology for each flow.

Business rules and Policies

A key role in the SOA promise of business agility and flexibility is by dynamic rules and policies.
These two seem to overlap each other and I got several questions by customers who drill into SOA architecture and concepts about what goes where - how should one determine whether a decision logic is a Policy or Rule?
I read a nice blog entry by Christina Lau, IBM describing this dilemma and giving some examples.