network latency?

[Joanne Sun]

Hello,

We are switching to distributed cache cluster (memcached) which is located on intranet and managed by another team.

But the latency increased from ms to s.

It seems the network latency since the cache is not local to the application. This is concerning because we use the cache cluster to cache the ML models stored on network disk. 

Wonder what I am missing here? Or I am doing some anti-pattern? What is the best practice of deployment of the distributed cache cluster? How to minimize the network latency?

Please advise. Thank you!

Joanne

[Toru Maesaka]

Hi there,

Request latency can be caused by so many different factors. I don't think any of us can provide definitive feedback based on the information you presented, but generally speaking, ms to s latency is very uncommon.

It would be very helpful if you could gather evidence of your network latency hypothesis. If the network is indeed saturated, I would then start questioning why it's saturated. Good first step would be to confirm if the network was always saturated, before memcached was even in the picture.

Cheers,

Toru

--

---
You received this message because you are subscribed to the Google Groups "memcached" group.
To unsubscribe from this group and stop receiving emails from it, send an email to memcached+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/memcached/c5cc5e6e-a3e2-45dc-927c-6320bc003359n%40googlegroups.com.