What happens to the specifications if you do the data entry manual?  I would enter the details manual just to compare and see if there is a difference.  Not sure why it would be that low.

enter manually is the same result, i first attempt is enter information manually, and the report result same as the proxy and VBR cores is very low. my second attempt is to import RVtools report and hope it work better but unfortunately it show up to 2 cores for the proxy only. i even try to remove the CDP and capacity tier, just to focus on local backup for single site 340 Instances, 100TB, Backup windows 8 Hours, GFS 14 Days, 16 Weeks, 12 Month and 7 Years. the result still the same it show only 4 core for vbr and 1 core for proxy. 

Does it show that it is an all-in-one VBR/Proxy server or are they separate?  Just asking as I have not used the calculators for a while.  Maybe ​@david.tosoff can assist here.

it show separate compute resources for VBR, Veeam DB, Veeam One, Veeam Enterprise,  proxy and repo server, but all with very little cores. 

Hi ​@kelvin koh,

The simplified math behind Veeam proxy sizing is this:

pData] x xChange Rate] x / /Backup Window] (excluding growth and some other factors)

If you take your 100TB example:

100TB x 5% (I’ve assumed) / 8h = 5TB/8h = 0.625TB/h

= 182 MB/s

= 182/ 80 per core (virtual) rounds out to around 2C

Based on BP proxy sizing estimates here (https://bp.veeam.com/vbr/2_Design_Structures/D_Veeam_Components/D_backup_proxies/vmware_proxies.html#processing-resources)

Now, as far as your 340TB example only equating to 2 cores, that’ll need more specifics for the unknowns to validate: If the change rate you’re using is low then that could be why. Likewise if compression is tuned off of default, etc.

Here is 340TB @ 1% vs 5%:
 

 my setup:

 Hi Kelvin, as linked above, the throughput rates are what affect the core count in the proxy math.

If you factor out the math manually, you’ll arrive at ~2.18 cores as well. Without physical proxy, you’ll arrive around 6.8 cores.

if based on manual calculation 340 Instance with 160 TB we should need 32 cores proxy as per advise from principle.

Hi Kelvin,

If you don’t mind, could you please export your scenario as a file and share it so we can have exact details about what inputs are used? 

Also, can you give an example of the manual math? I have presented the sizing courses at VeeamON the past several years and the processing rate used for manual formulas used to be 25 MB/s but now is variable based on the chart that ​@david.tosoff posted above. At minimum (80 MB/s), the proxy resources are reduced to approx. 33% of previous formulas. Best case (250MB/s) they are reduced to 1/10 of previous numbers. All of this based on extensive performance testing by our QA teams.