1. Cloud Hosting
  2. Regions
Regions

Regions & Latency: How to Pick Among Five Sites

Four sites across Asia-Pacific plus one on the US West Coast. Check measured median latency, model availability, and network specs—five minutes to find the Mac mini closest to you and the App Store.

Node Map

Five Regions, All Tier III+ Facilities

Each region runs an independently racked Mac mini cluster, online 365 days a year with no scheduled downtime windows. Facility tiers below.

Hong Kong

HK1
Facility tier
Tier III+
Latency to East Asia
32 ms

The top pick for East Asia. Premium backbone routing gives SSH sessions a near-LAN feel—great for daily remote coding and on-device testing.

Tokyo

JP1
Facility tier
Tier III+
Latency to East Asia
55 ms

All three tiers are available to order year-round. Balanced latency to East Asia, Korea, and the US West Coast—a solid middle ground for distributed teams.

Singapore

SG1
Facility tier
Tier III+
Latency to East Asia
88 ms

The natural access point for Southeast Asia-focused teams, with better latency to Jakarta, Bangkok, and Mumbai than any of our other five regions.

Seoul

KR1
Facility tier
Tier III+
Latency to East Asia
68 ms

Our newest APAC region, offering all three tiers—built for teams based in Korea or serving both Korean and Japanese markets.

Silicon Valley

SJC
Facility tier
Tier III+
Latency to US East
68 ms

The shortest path to App Store Connect uploads—fastlane deliver runs steadily at full bandwidth here. All three tiers in stock.

Availability

Tier × Region Availability Matrix

Check this table before you order. Each tier is offered in the regions shown below; the configurations actually available at order time are returned live by the portal.

Availability of three Mac mini tiers across five regions
Tier Hong Kong HK1 Tokyo JP1 Singapore SG1 Seoul KR1 Silicon Valley SJC
Mini SM4 · 16GB · 256GB · $19.9/day
Mini MM4 · 24GB · 512GB · $39.7/day
Mini ProM4 Pro · 64GB · 2TB · $60.5/day
In stock — delivered in ~5 minutes after ordering through the portal Not offered in this region

Full spec and daily/monthly pricing breakdown on the rental plans page.

Latency

Latency Reference: Median Ping Across Five Regions

Data reflects the median of hourly ICMP probes over the past 30 days, sampled from a mix of mainstream residential ISPs and datacenter uplinks in each region, in milliseconds. Your actual numbers will vary with your local ISP route—verify with the commands below before ordering.

Median ping latency from each region to major cities, in milliseconds
Region Shanghai Tokyo Seoul New York
Hong Kong HK1 32 ms 48 ms 62 ms 205 ms
Tokyo JP1 55 ms 2 ms 34 ms 158 ms
Singapore SG1 88 ms 68 ms 78 ms 232 ms
Seoul KR1 68 ms 35 ms 2 ms 172 ms
Silicon Valley SJC 152 ms 105 ms 128 ms 68 ms
latency-check — bash
# run from your machine; swap host for any node below
mtr -rwzbc 100 hk1.oncemini.com
ping -c 20 sjc.oncemini.com
# median of 100 probes ≈ the table above; jitter > 15ms means try another ISP route

Rule-of-thumb thresholds for SSH usability: ≤80 ms feels instant when typing, 80–150 ms is still comfortable to work in, and above 150 ms you'll want VNC or mosh instead. Test hosts are listed at the bottom of this page.

Decision Guide

Three Common Scenarios, Matched to a Region

Daily dev work from East Asia → Hong Kong HK1

32 ms median latency puts both SSH and VNC in the "feels instant" zone. Remote coding, on-device testing, and incremental Xcode builds feel about as close to a local Mac as you'll get. If you ever want a nearby alternative, Tokyo JP1 at 55 ms is still comfortably workable.

Shipping to App Store Connect → Silicon Valley SJC

App Store Connect's upload endpoints live in North America, so round trips from our West Coast regions land in the 20 ms range. The same 200 MB .ipa uploads 3–5x faster from the West Coast than from Asia-Pacific, and fastlane deliver timeouts basically disappear. If your pipeline handles both build and submission, the West Coast is the least friction.

Shared machine for a distributed team → Tokyo JP1

55 ms to Shanghai, 34 ms to Seoul, 105 ms to the West Coast—Tokyo is the most balanced of the five across every direction. If a team split between Asia and the US shares one machine for build shifts, Tokyo keeps latency workable for both sides, with all three tiers available to order year-round.

Not sure yet? Rent by the day first: starting at $19.9, run your real build scripts and mtr, then switch regions if it's not the right fit—snapshots are portable. See rental plans for details, and support for migration steps.
Network

Network Specs: Consistent Across All Five Regions

Network specs don't vary by tier or region—you never pay extra for bandwidth.

1 Gbps

Dedicated upload and download bandwidth per machine, not a shared pool. iperf3 tests consistently hit 940 Mbps+, with no throttling and no oversubscription.

Unmetered traffic

No bandwidth billing, no monthly cap, no throttling past a quota. Pull CI dependencies constantly, upload .ipa files, sync model weights—no need to track usage.

BGP multi-homed

Each region connects through 3+ upstream carriers with automatic BGP routing; a single upstream failure fails over in seconds without dropping your SSH session.

IPv4 + IPv6

One dedicated IPv4 address per machine, plus a free IPv6 /64 block. The default firewall template only opens SSH and VNC ports.

Facility

Facility Compliance & Power

Tier III+ and ISO 27001

All five regions run in Tier III or higher facilities holding ISO 27001 certification. Racks are individually locked behind two-factor access control, with 90 days of video footage retained. Your data never leaves the jurisdiction of the region you choose.

Dual power feeds, always on

Three-tier redundancy of A/B utility feeds, UPS, and diesel generators, with independent PDU monitoring per rack. Every region runs 365 days a year with no scheduled downtime windows—see our trust & security page for full SLA terms.

Sourced through official Apple channels

Every Mac mini is purchased new through Apple's official enterprise channels and racked sealed—never opened, modified, or built from used parts. Serial numbers are included with your delivery email so you can verify warranty status yourself.

Looking Glass

Test IPs & Speed Test Files

Test before you order: ping / mtr the test hostname, then wget a speed test file to see real throughput. Files are random byte streams available in 100MB and 1GB sizes, with no limit on how many times you run them.

Hong Kong HK1

  • ping targethk1.oncemini.com
  • 100MBhk1.oncemini.com/speed/100MB.bin
  • 1GBhk1.oncemini.com/speed/1GB.bin

Tokyo JP1

  • ping targetjp1.oncemini.com
  • 100MBjp1.oncemini.com/speed/100MB.bin
  • 1GBjp1.oncemini.com/speed/1GB.bin

Singapore SG1

  • ping targetsg1.oncemini.com
  • 100MBsg1.oncemini.com/speed/100MB.bin
  • 1GBsg1.oncemini.com/speed/1GB.bin

Seoul KR1

  • ping targetkr1.oncemini.com
  • 100MBkr1.oncemini.com/speed/100MB.bin
  • 1GBkr1.oncemini.com/speed/1GB.bin

Silicon Valley SJC

  • ping targetsjc.oncemini.com
  • 100MBsjc.oncemini.com/speed/100MB.bin
  • 1GBsjc.oncemini.com/speed/1GB.bin

Example: wget -O /dev/null https://hk1.oncemini.com/speed/100MB.bin. Speed test hosts are deployed separately from user machines, so results reflect the true performance of the 1 Gbps line on the machine you rent.

$ rent --region hk1 --chip m4 --days 1

Pick a Region, Get SSH Credentials in 5 Minutes

Choose your region and tier when ordering through the portal—the actual availability at order time is returned live by the portal.