Hi All, I ran several test runs for getting the correlation between 'event processing time' and 'TopK capacity' for Frequency implementation (transformer extension). I have put the stats for each test runs below (Graphs are event vs processing time). According to the results, it seems we need to decide the most optimal TopK capacity for this.
Common tests specs, Input event stream has 2 attributes (symbol and price). Symbol will be the selected attribute for frequency calculation. Client - 50 events within a for loop without a sleep time in between (to get the behaviour for maximum in-event stream rate). Calculating algorithm runs in a synchronised method. *TOPK_CAPACITY = 100000000* *Test Results,* *Average event processing time = 700.02 ms* *Processing time for all 50 events = 35001 ms* *[image: Inline image 2]* *TOPK_CAPACITY = 10000000* *Test Results,* *Average event processing time = 296.68 ms* *Processing time for all 50 events = 14834 ms* *[image: Inline image 3]* *TOPK_CAPACITY = 5000000* *Test Results,* *Average event processing time = 3.88 ms* *Processing time for all 50 events = 194 ms* *[image: Inline image 4]* *TOPK_CAPACITY = 1000000* *Test Results,* *Average event processing time = 0.98 ms* *Processing time for all 50 events = 49 ms* *[image: Inline image 5]* According to the results, there is significant correlation between event processing time and TopK capacity. (TopK capacity represents the number of *distinct* attribute values for frequency consideration). Please suggest a suitable TopK Capacity value for a production environment. Then we can continue more testings changing some other parameters to find out the relations and best possible values. BR *Asok Aravinda Perera* Software Engineer WSO2, Inc.;http://wso2.com/ <http://www.google.com/url?q=http%3A%2F%2Fwso2.com%2F&sa=D&sntz=1&usg=AFQjCNGJuLRux6KkJwXKVUCYOtEsNCmIAQ> lean.enterprise.middleware Mobile: +94722241032
_______________________________________________ Dev mailing list Dev@wso2.org http://wso2.org/cgi-bin/mailman/listinfo/dev