Hello,


Just to close this off I separated the latency and throughput measurements and 
also noticed the main reason for the flat growth: a slow node in the cluster. 
With all clusters in the same subnet, I am seeing the following latencies:



Latency (ms)

Entries (K)     50   75   90   95   99   100

5               3    3    3    4    9    93

50             10    12   13   14    23    46

500            74    77   83   89   108   119

10,000        1457  2010 2304 2756  3471  3476



Thanks all for the help.



Manuel



-----Original Message-----
From: vkulichenko [mailto:[email protected]]
Sent: 06 August 2016 06:20
To: [email protected]
Subject: RE: Ignite performance



Hi Manuel,



Are the results in the latest table in milliseconds? I don't see significant 
growth with the number of elements there, am I missing something? Actually your 
query calculates average for the whole data set which means that the cache is 
scanned, so I would expect that the latency will depend on the cache size. 
Makes sense?



-Val







--

View this message in context: 
http://apache-ignite-users.70518.x6.nabble.com/Ignite-performance-tp6703p6825.html

Sent from the Apache Ignite Users mailing list archive at Nabble.com.

Reply via email to