My Photo

Become a Fan

DailyMile

Google Ad Skyscraper

« The dangers of virtualization and datagrids | Main | New command line utility for observing grids in Extreme Scale V6.1.0.3 onwards »

August 07, 2008

Comments

Paul Anderson

Billy... I profiled it this morning... nice to see people are thinking about fast serialisation but it's not as fast as it could be.... quite a lot of waste. I'm working on a similar tech at the moment, it's more limited in function but probably 100x faster.

Christof Meerwald

Aren't there already enough standards for this? What about XDR, CDR, ASN1?

And why invent another interface definition language and not use (a aub/super-set of) IDL?

Paul Anderson

Christof, One thing that seems to have defined protocol marshalling is variety and choice. A question would be why do XDR, CNR and ASN1 exist as we already have RMI/IIOP. A possible reason in this case I suppose is performance.

The comments to this entry are closed.