Hello,
First of all, thank you for your contribution in helping the Tor Network grow.
I took a look at your family of relays. I think the main issue is that your relays are still fairly new on the Tor network so the usage isn't that high. It often takes months to see the growth in utilization and traffic.
I often found that in order to get higher utilization and traffic, it would require your servers to be located in EU or at least US East so you have good peering and routing to the rest of the relay servers.
Personally, I run most of my relays using VPSes rather than large dedicated servers because it gives me a lot of flexibility to cancel when things don't work out or scale easily when needed.
Regarding what you should do? I think you should downsize the hardware on your dedicated servers and look for providers in US East and EU that can offer both large amount of bandwidth preferably unmetered but also have excellent peering/connection to high Tor traffic providers like Hetzner, OVH. In terms of hardware, Unless, you are expecting enormous traffic, there is simply no need to have such a high spec dedicated server. I'm sure it is not cheap to run them. Hope this helps!
John
On Fri, Apr 4, 2025, at 2:00 AM, Tor at 1AEO via tor-relays wrote:
Part 3 - Looking for feedback on our Tor relay configuration decision:
Looking for guidance on optimizing server / bandwidth spending for Tor guard / middle relays across datacenters.
Our current setup: • All same cost: Three servers in different datacenters, each with 10 Gbps unmetered bandwidth. • Identical hardware/software: Each server has 128 threads, 512GB–768GB RAM, and the same software stack. • Diversity benefits: Each server hosts fewer than 10 other relays, enhancing overall network diversity. • Primary differences: Geographic location and upstream peering/routing.
Current Utilization: • Server 1: 2Gbps utilized (20% utilization) • Server 2: 6Gbps utilized (60% utilization) • Server 3: 8Gbps utilized (80% utilization)
Overall, using 16 Gbps out of 30 Gbps purchased -- roughly half our bandwidth is idle / "wasted". Assuming we can't significantly optimize these servers any further and given the significant under utilization on Server 1 (80% idle) and Server 2 (40% idle), what would you do in this scenario? Some options we’re considering: • Consolidation: Paying only for the servers with the highest utilization. • Repurposing: Allocating the idle bandwidth to another purpose that benefits Tor (or a similar effort). • Reinvestment: Using the funds from Servers 1 and 2 to acquire additional high-utilization capacity (e.g., more of Server 3 configuration) to boost our overall contribution closer to 24 Gbps. • Other strategies: Any alternative approaches you might suggest.
Appreciate any thoughts or best practices to help reduce wasted expenditure while maximizing the utility of our available resources.
Thanks for your insights!
tor-relays mailing list -- tor-relays@lists.torproject.org To unsubscribe send an email to tor-relays-leave@lists.torproject.org