Hi All, We have a requirement to do some calculations for an incoming tuple against each similar record in a database, assuming 200,000 such calculations. Each calculation is independent and has to be done real time. So a kind of 1 to many processing. What are the options that can be used involving Storm? One elementary approach is to decide a specific number of bolts say 5 and each bolt does 40000 intermediate calculations and then a final bolt can consolidate it.
Will Trident help in this area? Can we reduce/eliminate the sequential processing of 40000 calculations for example and make it parallel. What are the other better options? Thanks Anand
