How DNS and GSLB Enable Multi-Active Data Center Load Balancing

DNS Fundamentals

DNS is a distributed database that maps domain names to IP addresses, allowing users to access services without remembering numeric addresses. It operates as a hierarchical system with four main types of name servers: root, top‑level domain (TLD), authoritative, and local servers.

Typical DNS Resolution Flow

The client enters a domain (e.g., www.example.com) in a browser, triggering a DNS query to the local DNS resolver provided by the ISP.

The local resolver checks its cache; if the record is missing, it queries a root server.

The root server returns the address of the appropriate TLD server (e.g., .com).

The TLD server points the resolver to the domain’s authoritative name server.

The authoritative server finally returns the IP address, which the resolver caches for future requests.

TTL (Time‑To‑Live) and Caching

TTL defines how long a DNS record can stay in a resolver’s cache. For stable services, a long TTL (up to 24 hours) reduces query traffic and improves response time; for frequently changing services, a shorter TTL (e.g., 10 minutes on Alibaba Cloud DNS) ensures timely updates.

Multi‑Active Data Center Architecture

When two geographically separated data centers (A and B) serve the same application, DNS can influence which center a user reaches based on round‑trip time (RTT). By returning the IP of the nearest or fastest‑responding center, DNS helps achieve true multi‑active availability.

GSLB (Global Server Load Balancing) Example

Client sends a DNS request to its local resolver.

The resolver receives a list of four GSLB servers (e.g., DNS‑CTC, DNS‑CNC, DNS‑USA1, DNS‑USA2) from the root.

Local resolver polls these GSLB servers until one replies.

The responding GSLB server checks its proximity table; if a matching entry exists, it returns the IP of the closest data‑center web server.

If no entry exists, the GSLB queries another GSLB node to build the proximity information.

Both GSLB nodes probe the local resolver’s RTT (e.g., 50 ms vs. 100 ms) and record the faster path.

Based on the proximity table, the GSLB returns the optimal IP (e.g., 1.1.1.1) to the local resolver.

The resolver finally delivers this IP to the client, completing the request.

Load‑Balancing Strategies in Multi‑Active Environments

Geographic‑based routing: choose the data center closest to the client’s location.

RTT‑based routing: select the data center with the smallest round‑trip latency as measured by GSLB.

Both strategies can be implemented with dedicated hardware (e.g., F5 GTM) or with simple DNS round‑robin configurations for smaller workloads.