moebig
Member
OP
RAID is not backup
- Joined:
- Jul 2024
- Posts:
- 115
- From:
- Casablanca, Morocco
Wesh everyone. So I moved my blog to Vultr singapore region because "asia pacific growth incha'allah" and honestly... le server حلو, works fine. I am in marseille. My users are everywhere. With cloudflare in front, who cares if origin is SG or FR or US ?
I see people here obsessing over "must be in Frankfurt for EU users" or "tokyo for APAC". Bro. The CDN edge is incha'allah doing the work. Your origin could be on the moon if the cache hit ratio is good. The location fetish is outdated, wesh.
My setup : static site, cloudflare, singapore origin. Loads in 0.8s from paris. 0.9s from montreal. 1.1s from lagos. Prove me wrong. Location is a meme for most use cases.
RAID 1: because paranoia pays
armstrongvds
Member
ARM Enjoyer
- Joined:
- Jun 2024
- Posts:
- 199
- From:
- Seoul, KR
Moebig you are missing the edge case where it matters!
For dynamic workloads — game servers, real-time collab, financial tick data — you want that Ampere core as close to users as possible! Insane value from ARM at the edge, runs circles around x86 on watts.
I ran YABS (https://github.com/masonr/yet-another-bench-script) on:
- CloudCone Los Angeles: 18423 multi-core
- Same plan in Los Angeles: 18201 multi-core
But latency? 12ms vs 147ms! For my use case that is everything! Cache does nothing for WebSocket streams!
one small ping for man...
SingaporeRep
Member
Benchmark Addict
- Joined:
- Jul 2024
- Posts:
- 227
- From:
- Singapore, SG
# yabs.sh on moebig's "global" site via VPN exit in Jakarta
CPU: N/A (cloudflare edge)
Disk: N/A
Network:
Vultr Singapore: 8.2s TTFB, 2.1MB page
Same site via Frankfurt proxy: 0.8s TTFB
Decent theory, meh execution. Your singapore origin is getting hammered on cache miss. 8 seconds is not a meme, its a timeout.
1) check cache headers
2) move origin closer to your actual users
3) or admit hot take was lukewarm
fio, iperf, geekbench. results or gtfo.