Switching

we have 10 Gig interface receving  2 Gig of traffic

Also we have 1 aggregate interface containing 2 physical links ( each of them is 1 Gig )

We found the below issue :

The received 2 Gig traffic are not balanced to both links at the aggregate , the 2 Gig is sent to the first link of the aggregate

leading to many drops on it ( while the second link of the aggregate is no useda at all )

we confirmed that using : run monitor traffic interface X   ( checking the bps)

Juniper updated us that this is normal as the hashing value is the same for this 2 gig stream so it will be balnced only to one link even if the link is over utilized

we tested by putting 2  different  streams ( srcip1 + dstip 1  , srcip 2+dstip 2  )  each of them 1 gig , we found that both of them are balanced to the first link of the aggregate ( leading to the same issue above ) , juniper updated us that this means that hashing output for both streams  is the same

How come that the interfca utilization is not take into consideration, how to change this behaviuor to balance on both links ?

Also how will the customer be sure that no drops will happen ?

This is normal behavior and in order to effect a more even distribution you need to have more than just a few flows.  Basically given your two flows, the result of the two hashes is producing the same output.  You should try more than just two flows to simulate a more real-world environment where there are typically hundreds to thousands of flows traversing a link at any given time.

Thanks