Hacker Newsnew | past | comments | ask | show | jobs | submit | Hardops's commentslogin

No GPUs, yet.

Yes, the 3TB is 2/3 used for Caching, and 1/3 as a local storage for our SSL/DNS/Firewall/etc.


Is the 2/3 for caching using part of each NVMe, or is there one dedicated for the rest? Mostly wondering about spreading writes more evenly, rather than peak throughout.


hey, this is Rami ... I lead HW team at Cloudflare. Great Q! .. in our case, there was a sweet spot between the number of cores, the L3 cache per core, NUMA latencies, mem bw per core, cost & power. The 48c gave us the best req per second per $, and it likely was due to a combination of all these things.

We have two more blogs that will come out probably today, which will shed more light on why AMD worked better for us.


Hi, this is Drew from Netflix.. Just curious, are you running single socket rome in numa or non-numa mode? I've seen better performance in NPS=4 mode myself. We seem to see about a 7% increase when moving from NPS=1 to NPS=2 and another 7% going to NPS=4.


Yes, we got the most perf when we used NPS=4.

Stay tuned for 3 more blogs this week ... they will give deeper analysis on the perf % gains we saw and why.


This is Rami from Cloudflare ... we actually use DAC Cables w/ SFP+ for all of our 25G nics.


hey, this is Rami .. I lead hw team in Cloudflare.

Great observation! these chassis are designed for 19" racks, not OCP spec. But, who does not love green thumb screws!


Can I ask the chassis and motherboard combination? I know previous generations have been based on QCT chassis and motherboard components. I am curious what the non-hotswap, shallow depth, 1S is as it probably saves a fair amount of money dropping the HPC/density and hot swap premiums.


We're still using ODM built servers that are available for the public. Yes, moving off the shared infra chassis helps us improve reliability & cost thru simplicity, and enables us to scale our infrastructure using single nodes vs. X nodes as the smallest scaling unit. Our machines are stateless as our globally interconnected infra is widely distributed and replicated. Hence we moved to no-hotplug.


Looks like an OCP mezzanine interface tho.

If a server can take OCP mezz cards and the BMC can be managed in the same manner and with the same tools as a fully OCP-spec system, does it really matter that it's not in an OCP-spec chassis and rack?


Yes, the chassis form factor does not matter much to us at this point.

Yes, we're using OCP Nic ... and many other components that are used by other servers/clouds. Nothing proprietary except our SW stack.


Consider applying for YC's Summer 2026 batch! Applications are open till May 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: