[dpdk-dev] DPDK ACL Trie

[dpdk-dev] DPDK ACL Trie

[Gowda, Deepak]

Hi,

I'm experimenting on DPDK trie for ACL implementation. It turns out the more the number of rules, the more time rte_acl_build() takes which is expected. But, I was wondering if there are any standard benchmarking data that indicates build time for 'x' number of rules.
I benchmarked it myself, the build time for a few hundred rules is quite large really. It's in the order of secs, I can't help but think how could this happen, I mean in networking we typically can't afford 'seconds' of delay right.?
I just want to recheck my data against yours (if any) just to be sure.
With trie build-time being the constraint, I'm unable to scale up the no. of ACLs.

Hoping to see my query answered.


Thanks,
Deepak Gowda


-----------------------------------------------------------------------------------------------------------------------
Notice: This e-mail together with any attachments may contain information of Ribbon Communications Inc. that
is confidential and/or proprietary for the sole use of the intended recipient.  Any review, disclosure, reliance or
distribution by others or forwarding without express permission is strictly prohibited.  If you are not the intended
recipient, please notify the sender immediately and then delete all copies, including any attachments.
-----------------------------------------------------------------------------------------------------------------------