Choose how users connect
Use an Anycast endpoint for a single global entry point, or choose individual regions for more control over where users enter the network.
LuckVM Global Network Accelerator LGA
Keep your origin. Extend your reach.
Connect distributed users to your APIs, remote desktops and TCP/UDP services through a managed acceleration endpoint. Keep your existing public origin, choose a routing model and manage listeners, TLS and outbound usage in one place.
Endpoint β Accelerated path β Origin
BUILT FOR CONNECTIONS
Global acceleration is designed for connections that must reach an origin: dynamic requests, database sessions and interactive remote access. It complements your deployment without requiring a server migration.
Use an Anycast endpoint for a single global entry point, or choose individual regions for more control over where users enter the network.
Configure listener and origin ports for TCP/UDP workloads. Use TCP passthrough for origin-side TLS, or bind a certificate for an HTTPS listener.
Review endpoint addresses, deployment status and outbound usage. Update primary and backup origins, listeners and health checks in the customer portal.
TWO WAYS TO CONNECT
Choose according to user distribution, required access regions and the IP allowance included in your plan.
Use an Anycast IP for global access without selecting regions individually. Availability depends on your origin and plan.
Use intelligent DNS and dedicated regional IPs to direct traffic into selected regions. Optional Mainland China acceleration makes eligible CN2 routes available.
CHOOSE YOUR WORKLOAD
Choose a preset, then adapt the listener ports and routing to your actual workload.
Connect applications and database clients across regions with TCP listeners for services such as MySQL and PostgreSQL.
TCP Β· Database clientsConfigure SSH and RDP access for distributed operations teams. Remove unused preset ports and retain your normal access controls.
SSH Β· RDP Β· Remote administrationAdd an acceleration path to an existing domain-based architecture. Choose TLS passthrough or a certificate-bound HTTPS listener and avoid origin loops.
Hostname endpoints Β· TLSConfigure ports and routing for APIs, interactive applications and other services that need to reach a live origin.
TCP / UDP Β· Custom listenersPLANS FOR YOUR BUSINESS
Compare included outbound transfer, acceleration IPs, concurrent connections and overage rates. Select the subscription term when you configure your service.
Match the allowance and connection limits to your expected workload.
Loading available plansβ¦
The checkout page shows the final amount. Outbound allowance resets monthly; Mainland China and other routes have separate overage rates.
FROM PLAN TO CONNECTION
Prepare a reachable origin, its service ports and DNS access if you use a hostname endpoint.
Sign in and select the allowance and optional Mainland China access you need.
Choose the term, origin region, routing and listener-to-origin port mappings.
Open your instance in the customer portal and confirm the deployment status.
Use the acceleration IP or set your CNAME record. Test from real user networks.
Allow the required origin ports and test end-to-end connectivity. A hostname endpoint must differ from its origin hostname. For HTTPS, install the certificate on the origin for TCP passthrough or upload it in the portal for TLS termination.
FIND THE RIGHT FIT
Choose for your traffic pattern. The two can also be used together when the request path is configured correctly.
| Requirement | LuckVM LGA | CDN |
|---|---|---|
| Primary focus | Connection and transport paths to an origin | Caching and edge content delivery |
| Typical workloads | Dynamic APIs, remote access, TCP/UDP services | Cacheable images, scripts and downloads |
| Configuration | Origin, routing and listeners | Hostname, cache policy and origin settings |
| Using both | Validate the complete request path and avoid circular origin references. | |
BEFORE YOU START
Clear answers about billing, Mainland China access, TLS and what acceleration can realistically improve.
No server migration is normally required: LGA connects users to your existing public origin. It optimizes the connection path rather than providing application hosting or a general-purpose content cache. A CDN remains useful for cacheable assets; validate the request path before combining the two.
No. Enable Mainland China acceleration when comparing or configuring plans, then choose custom routing and an available CN2 region. Standard Anycast routing does not by itself select a CN2 route. The Mainland China option has its own price and outbound overage rate.
Only outbound traffic from the acceleration network to users is counted; inbound traffic is not charged. The allowance resets on the monthly anniversary of activation, using Asia/Shanghai time, rather than every 30 days. If the anniversary date does not exist in a month, the boundary falls on that monthβs last day.
With outbound overage balance billing enabled, usage beyond the allowance is charged at your planβs rate. A detected balance of zero or less triggers suspension. Without balance billing, exhausting the allowance triggers suspension. Metering and shutdown take time; additional usage is still chargeable and may make the balance negative.
For TCP 443 passthrough, keep the certificate on your origin. To terminate TLS at the acceleration edge, use a hostname accelerator and upload its matching PEM certificate chain and private key in the customer portal. After binding finishes deploying, configure the HTTPS listener and origin protocol.
Standard routing uses one acceleration IP. With custom routing, each selected region uses an acceleration IP and the maximum number is set by your plan. Available and disabled regions depend on the origin and Mainland China option; use the live configuration list.
There is no universal latency reduction percentage. Results depend on user networks, origin location, routing, protocol and server performance. Compare direct and accelerated connections from representative user networks before moving production traffic.
Check the instance status in the customer portal and contact support with the instance details. Do not place a duplicate order just to retry provisioning. Deployment is asynchronous, so a submitted request is not proof that the endpoint is ready.
Choose a plan, or discuss your origin and user regions with our team.