seeing some inconsitencies in our DMIs when we add the Link Group dimension.
Link Group, our Total Bytes for a given transaction is ~9.5GB. When Link Group
is added, that same Transaction sees >10GB of traffic. Are the CAS's somehow
removing duplicate data? I thought that'd be done at the AMD level.
Solved! Go to Solution.
Yes the CAS server has a de-duplication algorithm if multiple AMDs see the same data. If you add link information it forces it to show you all data.
AMD de-duplication is if it sees the same packet multiple times. An AMD has no visability of what any other AMD sees.
Hi Alasdair, I'm not convinced this algorithm is working correctly, as the discrepancy between traffic not split by link group and traffic that is seems rather high. The below image is for one Software service, both un-split and split. Any advice you have would be most greatly appreciated.
Thanks for your response! I didn't know the CASs were capable of deduplication - good to know!
With my Stores 'Service Protocol', the SW-C Link Group has 1.48 TB of bytes. Even if all the other Link Group data was subtracted, the Total Bytes would be 1.47 TB. Any idea why that's reported so differently than without Link Groups? The SW-C Link Group AMDs all report to the same CAS/ADS cluster.