Compute Providers & Regions
Choose the right provider and region for your bots based on cost, latency, and exchange proximity.
Overview
Every bot runs on a dedicated compute instance — no shared resources, no noisy neighbours. When you deploy a bot, you select a provider and region. The instance is provisioned automatically and torn down when you stop the bot.
plutarc currently supports three providers across EU, US, and Asia-Pacific data centres. All providers deliver the same bot functionality — the difference is pricing and proximity to exchange matching engines.
Providers
Hetzner (CCX23)
Price range: £0.13–0.24/hr
Vultr (VOC-G-4C-16GB-80S)
Price range: £0.56–0.62/hr
AWS (M6I.XLARGE)
Price range: £0.68–0.78/hr
See the Pricing page for the full per-region breakdown.
Choosing a Region
The most important factor is proximity to the exchange matching engine. Deploying in the same city as the matching engine minimises round-trip latency for order placement, fills, and position updates.
Exchange Matching Engine Locations
| Exchange | Matching Engine |
|---|---|
| BitMEX | Tokyo |
| Bybit | Singapore |
| Binance | Tokyo |
| Kraken | London |
Datacenter locations are based on publicly available information and may change without notice.
Latency by Region
Approximate round-trip time from each deployment region to exchange API servers, measured periodically by probes in each datacenter.
| Region | BitMEX | Bybit | Binance | Kraken |
|---|---|---|---|---|
Helsinki, Finland Northern EU · hetzner | ~341ms | ~228ms | ~296ms | ~156ms |
Falkenstein, Germany Central EU · hetzner | ~314ms | ~215ms | ~287ms | ~106ms |
Dublin, Ireland Western EU (AWS) · AWS | ~271ms | ~209ms | ~236ms | ~76ms |
London, UK Western EU (AWS) · AWS | ~247ms | ~192ms | ~243ms | ~43ms |
London, UK Western EU (Vultr) · vultr | ~302ms | ~230ms | ~299ms | ~72ms |
Nuremberg, Germany Central EU · hetzner | ~300ms | ~189ms | ~273ms | ~82ms |
Ashburn, USA US East Coast · hetzner | ~235ms | N/A | N/A | ~150ms |
Hillsboro, USA US West Coast · hetzner | ~164ms | N/A | N/A | ~273ms |
Singapore Southeast Asia · hetzner | ~146ms | ~46ms | ~128ms | ~350ms |
Singapore Southeast Asia (Vultr) · vultr | ~151ms | ~62ms | ~135ms | ~435ms |
Singapore Southeast Asia (AWS) · AWS | ~122ms | ~31ms | ~102ms | ~360ms |
Tokyo, Japan Asia-Pacific (Vultr) · vultr | ~69ms | ~121ms | ~79ms | ~513ms |
Tokyo, Japan Asia-Pacific (AWS) · AWS | ~51ms | ~97ms | ~27ms | ~307ms |
N/A indicates the exchange does not support API access from that region.
Green (<50ms) = same geographic area as matching engine. Amber (50–150ms) = adjacent region. Red (>150ms) = cross-continental.
When Latency Matters
Not every strategy needs the lowest possible latency. Consider your timeframe:
- Short timeframes (1m, 5m) — Latency has a direct impact on fill quality. Deploy in the same city as the exchange matching engine for the best results. Vultr and AWS APAC regions are recommended for BitMEX, Bybit, and Binance.
- Medium timeframes (15m, 1h) — Adjacent regions (amber latency) are generally fine. The difference between 30ms and 100ms is negligible when candles close every 15–60 minutes.
- Long timeframes (4h, 1d) — Region choice has minimal impact on execution quality. Hetzner EU regions offer the best value.
Cost vs Latency
There is a clear trade-off between cost and proximity to APAC exchange matching engines:
- Budget-friendly — Hetzner EU (£0.13/hr). Cross-continental latency to APAC exchanges, but excellent for longer timeframes or EU-based exchanges like Kraken.
- Balanced — Hetzner Singapore (£0.24/hr) or Vultr APAC (£0.56/hr). Good latency to Bybit and Binance at moderate cost.
- Lowest latency — AWS or Vultr Tokyo/Singapore (£0.56–0.78/hr). Best for latency-sensitive strategies trading on 1m–5m candles.
All providers deliver the same bot functionality, strategy evaluation, and risk management. The only differences are pricing and network latency to the exchange.