๐ฑ๐ฐ Serving from Colombo ยท Sri Lanka
Your users are here.
Your servers should be too.
Most Sri Lankan apps run their infrastructure oceans away, and every request pays the crossing twice. Roar AI Cloud puts your app, your database and your AI gateway in one Colombo facility. Everything below is measured live โ from your browser, from our servers, right now.
Your round trip ยท measured in your browser
Colombo edge ยท 0 samples, one every 0.75s ยท warm-up excluded
- fastest
- โโ ms
- median
- โโ ms
- p95
- โโ ms
Ten calls, three regions.
Ten identical calls to each region, in sequence over one connection โ the way an application really talks to a gateway. What separates them is whether the request leaves the building, not how far it travels.
AI gateway ยท 10 calls per region
Run from this server, one connection per region, calls in sequence โ the way an application actually uses a gateway.
- Colombosame facilityโโ
measuringโฆ
- Singaporeover the internetโโ
measuringโฆ
- Bangaloreover the internetโโ
measuringโฆ
Colombo is faster because the gateway sits in the same facility as your app, so the request never crosses the internet. Singapore and Bangalore land within a similar range of each other despite the difference in distance โ Sri Lankan transit reaches both over much the same submarine route, so kilometres are not what separates them.
Try it yourself.
We deployed this exact application to Railway as well. Run the test and your own browser will time both โ no figures from us involved.
Same application ยท two platforms
The identical build runs here in Colombo and on Railway in Singapore. Your browser calls both, 8 times each, and reports what it measured.
Network round trip
A request that does almost nothing, so this is the connection itself.
ColomboโโRailway ยท SingaporeโโDatabase-backed request
A request that queries Postgres and returns rows โ a real page's work.
ColomboโโRailway ยท Singaporeโโ
Railway has no Sri Lanka region โ Singapore is the closest they offer, so this is their best case rather than a poor configuration. The gap is distance, not a deficiency. Measured from wherever you are: if you are reading this outside Sri Lanka the result will differ, and it should.
Writes that land now.
A dedicated Postgres container beside your app, on the same private network. Press the button and watch real rows commit.
Dedicated Postgres ยท real committed writes
Your own container, not a shared cluster.
fastest write
โโ ms
median write
โโ ms
rows / second
โโ
read back
โโ ms
Three regions, one platform.
Colombo
Sri Lanka
- app hosting
- available
- postgres
- dedicated
- ai gateway
- api.lk.roar-ai.com
- from colombo
- โ
Singapore
Singapore
- app hosting
- available
- postgres
- dedicated
- ai gateway
- api.sg.roar-ai.com
- from colombo
- 2,740 km
Bangalore
India
- app hosting
- available
- postgres
- dedicated
- ai gateway
- api.in.roar-ai.com
- from colombo
- 1,300 km
Every number on this page was measured while you read it.
No cached benchmarks, no vendor-supplied figures. Reload and they change, because they are real requests over your connection and ours.
Send a prompt through Colombo โ