Nearest to
Malaysia, Indonesia, Thailand, Vietnam and the Philippines, all a short hop across the region. It is also the nearest of the three to Australia and East Asia.
VPS and VDS in Singapore, for users and customers in Southeast Asia. The same prices, processor, port and filter as every other region.
| To | Distance | Bearing | Round-trip floor |
|---|---|---|---|
| Mumbai | 3,903 km | 302° | 39.0 ms |
| Limburg | 10,300 km | 321° | 103.0 ms |
Great-circle distance from the coordinates. Floor = 2 x distance / 200,000 km/s, light in fibre there and back: physics, not a measurement. Measured round trips are always higher.
Measured: a ping between our own Singapore and Mumbai servers on 1 October 2026 averaged 61.4 ms (lowest 60.5, highest 65.9), 1.57 times the 39.0 ms floor. One path on one day between two of our servers, not a promise for your connection.
| Plan | VCPU | Memory | NVMe | Traffic | Per month | Order | |
|---|---|---|---|---|---|---|---|
| NanoA VPN, a Discord bot, a small website or a lab box. | 1 | 2 GB | 20 GB | 1 TB | ₹890$9.49 | Deploy | 1 vCPU, 2 GB, 20 GB NVMe, 1 TBA VPN, a Discord bot, a small website or a lab box. |
| MiniAn API, a staging copy of production, or a small database. | 2 | 4 GB | 40 GB | 2 TB | ₹1,780$18.99 | Deploy | 2 vCPU, 4 GB, 40 GB NVMe, 2 TBAn API, a staging copy of production, or a small database. |
| StarterA production web app with its database on the same box. | 3 | 6 GB | 60 GB | 4 TB | ₹2,670$27.99 | Deploy | 3 vCPU, 6 GB, 60 GB NVMe, 4 TBA production web app with its database on the same box. |
| GrowthRecommendedA busy web shop or SaaS back end, with its workers and cache alongside. | 4 | 8 GB | 80 GB | 6 TB | ₹3,560$37.49 | Deploy | 4 vCPU, 8 GB, 80 GB NVMe, 6 TBA busy web shop or SaaS back end, with its workers and cache alongside. |
| ProCI runners, several services side by side, or a busy game server. | 6 | 12 GB | 120 GB | 8 TB | ₹5,340$56.49 | Deploy | 6 vCPU, 12 GB, 120 GB NVMe, 8 TBCI runners, several services side by side, or a busy game server. |
| EliteDatabases that want memory, and memory-hungry application servers. | 8 | 16 GB | 180 GB | 10 TB | ₹7,120$74.99 | Deploy | 8 vCPU, 16 GB, 180 GB NVMe, 10 TBDatabases that want memory, and memory-hungry application servers. |
| MaxHeavy compute: build servers, render queues, analytics jobs. | 10 | 20 GB | 260 GB | 14 TB | ₹8,900$93.99 | Deploy | 10 vCPU, 20 GB, 260 GB NVMe, 14 TBHeavy compute: build servers, render queues, analytics jobs. |
| TitanThe largest VPS: when you want most of a machine without a VDS. | 12 | 24 GB | 400 GB | 18 TB | ₹10,680$112.49 | Deploy | 12 vCPU, 24 GB, 400 GB NVMe, 18 TBThe largest VPS: when you want most of a machine without a VDS. |
Every size works out to ₹890 on every sizebetween $9.33 and $9.49 per vCPU a month: the monthly price divided by the vCPU count, rounded. Traffic is per month.
| Plan | Cores | Memory | NVMe | Traffic | Per month | Order | |
|---|---|---|---|---|---|---|---|
| MicroCoreOne latency-sensitive service that should never wait for a shared core. | 2 | 8 GB | 100 GB | 8 TB | ₹4,970$52.49 | Deploy | 2 cores, 8 GB, 100 GB NVMe, 8 TBOne latency-sensitive service that should never wait for a shared core. |
| FusionCoreRecommendedA production database with steady load, or a large game server. | 4 | 16 GB | 200 GB | 10 TB | ₹9,940$104.49 | Deploy | 4 cores, 16 GB, 200 GB NVMe, 10 TBA production database with steady load, or a large game server. |
| OmniCoreSeveral production services on one machine, each with cores of its own. | 6 | 24 GB | 500 GB | 12 TB | ₹14,910$156.99 | Deploy | 6 cores, 24 GB, 500 GB NVMe, 12 TBSeveral production services on one machine, each with cores of its own. |
| PrimeCoreBusy SaaS back ends and hosting resellers. | 8 | 32 GB | 800 GB | 16 TB | ₹19,880$209.49 | Deploy | 8 cores, 32 GB, 800 GB NVMe, 16 TBBusy SaaS back ends and hosting resellers. |
| TitanCoreMemory-heavy workloads: large databases, caches, search. | 12 | 48 GB | 1 TB | 24 TB | ₹29,820$313.99 | Deploy | 12 cores, 48 GB, 1 TB NVMe, 24 TBMemory-heavy workloads: large databases, caches, search. |
| UltraCoreAll sixteen cores of the processor, allocated to you. | 16 | 64 GB | 1.5 TB | 30 TB | ₹39,761$418.99 | Deploy | 16 cores, 64 GB, 1.5 TB NVMe, 30 TBAll sixteen cores of the processor, allocated to you. |
Every size works out to ₹2,485 on every sizebetween $26.12 and $26.25 per core a month: the monthly price divided by the core count, rounded. Traffic is per month.
Every plan: AMD EPYC 4584PX, up to 5.7 GHz, DDR5, NVMe, a 1 Gbit/s port, KVM, full root / Administrator access, DDoS protection included, always on. Bare metal in Singapore is quoted per build: see bare metal.
Attacks big enough to fill the link are removed upstream by 17 Tbps of mitigation capacity in front of the network. It is the network's figure, not something Clouduxe runs, and it is not metered or divided per customer. What is left reaches the host's 25 Gbit/s network card, where Prism checks it before your server's 1 Gbit/s port.
Clouduxe Prism, our XDP/eBPF filter, runs at the driver of the host's 25 Gbit/s card and judges each packet before your server's port or operating system sees it. You set its port rules and packet-rate ceilings on the Firewall tab; the detection engines run on every open port.
Traffic for FiveM, Minecraft, Source, rAthena, TeamSpeak, RDP, DNS and VPN is checked against the service's own protocol, so a flood that looks like real clients but never speaks it is dropped. Rust, SA-MP / open.mp and ARK: Survival have one-click port presets with rate limits and the detection engines, without protocol validation.
Nano and UltraCore sit behind the same upstream capacity and the same filter, in Mumbai, Singapore and Limburg alike, on from the first boot.
Comparing hosts? How to check a host's DDoS protection claims. In the help center: how protection works and protocol validation.
Upstream, in front of the network
17 Tbps
Mitigation capacity that absorbs floods big enough to fill the link. The network's figure, the same for every plan, not operated by Clouduxe.
Clouduxe Prism, on the host's card
25 Gbit/s
XDP/eBPF at the driver of the host's network card, before any server's port. Each open port is checked as deep as the rows below.
Your server's port
1 Gbit/s
On every plan in Singapore. Only what Prism passes arrives here.
Preset: TCP 30120, UDP 30120
Checked at layers 3, 4 and 7: The Cfx.re game and listing traffic
Preset: Java TCP 25565; Bedrock UDP 19132
Checked at layers 3, 4 and 7: Java edition, and Bedrock over RakNet
Preset: UDP 27015
Checked at layers 3, 4 and 7: CS2, CS:GO, TF2, Garry's Mod and the shared query protocol
rAthena
Checked at layers 3, 4 and 7: Ragnarok Online logins and zone traffic
Preset: UDP 9987, TCP 10011, TCP 30033
Checked at layers 3, 4 and 7: Voice, server query and file transfer
RDP
Checked at layers 3, 4 and 7: Remote Desktop
DNS
Checked at layers 3, 4 and 7: Query and response validation
VPN
Checked at layers 3, 4 and 7: Common tunnel protocols
Preset: UDP 28015, TCP 28016
Checked at layers 3 and 4: A port preset, rate limits and the detection engines. No protocol validator.
Preset: UDP 7777
Checked at layers 3 and 4: A port preset, rate limits and the detection engines. No protocol validator.
Preset: UDP 7777, UDP 27015
Checked at layers 3 and 4: A port preset, rate limits and the detection engines. No protocol validator.
Web server
Preset: TCP 80, TCP 443
Checked at layers 3 and 4: HTTP checks for request floods, missing client identity and off-path requests. Not a protocol validator.
Any other service
Checked at layers 3 and 4: The port rules and packet-rate ceilings you set, and the detection engines.
L3 is the address, L4 the transport (where your port rules' packet-rate ceilings apply), L7 the protocol. Each layer runs only if the one before did not decide. A port with a preset opens in one click from the Firewall tab; protocol validation is a switch on the port's rule.
Malaysia, Indonesia, Thailand, Vietnam and the Philippines, all a short hop across the region. It is also the nearest of the three to Australia and East Asia.
Services whose users are spread across Southeast Asia, where no single country dominates the audience.
Users in Europe: Limburg is the furthest region from Singapore on the corridor.
4 vCPU, 8 GB DDR5, 80 GB NVMe at ₹3,560$37.49 a month, for most web apps, APIs and services.
4 dedicated cores, 16 GB DDR5, 200 GB NVMe at ₹9,940$104.49 a month, for work that must not share a core.
Singapore adds 12.4 ms from Kolkata up to 37.0 ms from Ahmedabad to the round-trip floor, because it is 1,235 to 3,703 km further away.
From Kuala Lumpur, Jakarta, Bangkok, Ho Chi Minh City and Manila, Mumbai is the one that adds: 15.7 ms from Bangkok up to 37.5 ms from Jakarta.
Put it where most of your users are. If the audience is split, weigh the floor the smaller group would pay. Every plan costs the same in Mumbai and Singapore, so price never tips it either way.
How distance turns into latency, and once a server is running, diagnosing latency in the help center.
From cities in India, Mumbai is nearer
Kolkata
Mumbai 16.5 msSingapore 28.9 ms
Singapore adds +12.4 ms
Chennai
Mumbai 10.3 msSingapore 29.0 ms
Singapore adds +18.7 ms
Bengaluru
Mumbai 8.5 msSingapore 31.6 ms
Singapore adds +23.1 ms
Hyderabad
Mumbai 6.2 msSingapore 32.9 ms
Singapore adds +26.7 ms
Delhi
Mumbai 11.5 msSingapore 41.4 ms
Singapore adds +29.9 ms
Ahmedabad
Mumbai 4.4 msSingapore 41.4 ms
Singapore adds +37.0 ms
From Southeast Asia, Singapore is nearer
Bangkok
Singapore 14.3 msMumbai 30.0 ms
Mumbai adds +15.7 ms
Ho Chi Minh City
Singapore 11.0 msMumbai 37.3 ms
Mumbai adds +26.3 ms
Manila
Singapore 23.9 msMumbai 51.3 ms
Mumbai adds +27.4 ms
Kuala Lumpur
Singapore 3.1 msMumbai 36.0 ms
Mumbai adds +32.9 ms
Jakarta
Singapore 9.1 msMumbai 46.6 ms
Mumbai adds +37.5 ms
The floor is the great-circle distance between city centres, there and back at 200,000 km/s, the speed of light in optical fibre. It is the least a round trip can take, not a measured or promised ping: real routes are longer, and the equipment along them adds time. The added figure is the farther region's floor minus the nearer one's.
For communities across Southeast Asia and Australia
Java and Bedrock, for players across Asia-Pacific
Windows Server over Remote Desktop, in Singapore
Every game in the calculator: how its traffic is protected and the plan it starts on
The monthly cost by player count, worked out from our plans
Sites, shops and web apps, sized from your traffic
Yes. Singapore is the nearest of our three regions to all 13 Asia-Pacific cities on our list, from Kuala Lumpur at 308 km to Auckland at 8,418 km. Every Deploy button on this page orders a server in Singapore.
Yes, on every VPS and VDS plan, always on, with nothing to add at checkout. Floods big enough to fill the link are absorbed upstream by 17 Tbps of mitigation capacity in front of the network, the same for every plan. What is left meets Clouduxe Prism, our XDP/eBPF filter, at the driver of the host's 25 Gbit/s network card, before the server's 1 Gbit/s port; it validates the protocol for FiveM, Minecraft, Source, rAthena, TeamSpeak, RDP, DNS and VPN traffic. Rust, SA-MP / open.mp and ARK: Survival get port presets and rate limits, without protocol validation.
Mumbai, by distance, from every Indian city on our list: Singapore adds 12.4 ms (Kolkata) to 37.0 ms (Ahmedabad) to the round-trip floor. Singapore makes sense when much of your audience is in Southeast Asia, where Mumbai adds 15.7 ms (Bangkok) to 37.5 ms (Jakarta) instead.
It is the nearest of our three regions: Perth is 3,914 km from Singapore, Melbourne 6,063 km and Sydney 6,307 km, against 7,290 km to 10,157 km from the same cities to Mumbai. The round-trip floor is 39.1 ms from Perth and 63.1 ms from Sydney, before any routing.
No. Every plan costs the same in Mumbai, Singapore and Limburg: Growth, where we would start most web apps, is ₹3,560 or $37.49 a month in each, excluding GST.
Each server has a 1 Gbit/s port, and every plan includes monthly traffic, from 1 TB on Nano to 30 TB on UltraCore. The 17 Tbps figure is upstream mitigation capacity in front of the network, not the speed of your port.
Not from the panel. The region is part of the product you order and stays fixed for that server, and upgrades only offer larger plans in the same region. Moving it to Mumbai or Limburg goes through a support ticket, so pick the region nearest most of your users before you order.
8 VPS and 6 VDS sizes at the same prices as every region, behind 17 Tbps upstream and Clouduxe Prism, with a 99.99% SLA.