What is global server load balancing (GSLB)?

The primary goal of global server load balancing is to enhance application performance, reliability and availability. GSLB makes sure that user requests are routed to the best possible data center based on factors like server load, proximity to the user and network conditions.

Global load balancing differs from standard load balancing, which balances local traffic and is a standard requirement for all enterprises. Standard load balancing prevents specific back-end resources from being overloaded and optimizes performance across local servers and services within a single cloud or on-premises data center.

GSLB provides load balancing for load balancers, typically those operating in large, distributed enterprises with massive global footprints that manage high-volume data throughput. It prevents local and regional load balancers that operate in multicloud environments—and those orchestrating workloads between clouds and on-premises infrastructures—from being overloaded, optimizing performance across global servers and services.

GSLB is a vital part of maintaining high application availability, especially for large-scale web services and enterprise networks. It relies on a few key processes and components to function.

GSLB often relies on manipulating the domain name system (DNS) to make intelligent routing decisions. When a user requests a web address or service that is distributed across multiple locations, the user’s device initiates a DNS query that is intercepted by a GSLB-enabled DNS server or service.

The GSLB system directs the user query to the IP address with the most suitable server location. Location determination is based on a set of policies, the current state of the network and with the help of advanced geolocation technologies when needed. For example, user requests initiated in Mexico are routed to DNS servers in Mexico while user requests from New Zealand are directed to servers located in New Zealand, or the next nearest location.

ADCs are the brain of any GSLB system: these specialized devices or software components collect data on server health and traffic patterns to make real-time decisions on how to distribute incoming requests. The decision-making process is dynamic, so the ADC’s decisions can change as conditions fluctuate.

GSLB systems perform regular health checks on all servers in the pool to ensure they're functioning optimally and facilitating a seamless user experience. Health checks can include simple pings, TCP connections or more sophisticated application-level transaction testing. If a server fails a health check, the GSLB system can remove it from the rotation until it’s up and running again.

GSLB services continually measure the round-trip time (RTT)—the time it takes for a data packet to go from client to server and back again—to help the system continually improve its routing decisions.

Global server load balancing services rely on a range of algorithms, like geolocation-based routing, to distribute network traffic efficiently.

Round-robin load balancing, for instance, distributes requests evenly across a set of servers, while least-connections load balancing directs traffic to servers with the fewest active connections. Least-time load balancing sends user requests to the server with the fastest response time, and weighted load balancing assigns requests based predetermined weights that indicate server capacity.

With anycast, a complex network routing method, multiple servers share the same IP address, allowing clients to connect to a single address. With the help of border gateway protocol (BGP) routing mechanisms—which advertise the same IP prefix from different locations—incoming client queries are then routed to the nearest server (in terms of network topology or least-time delay).

Anycast can also improve network resilience, as traffic is automatically rerouted to the best available path. If a server outage occurs, requests are routed to the next-nearest data center so that, even if entire data centers go offline, users experience a small reduction in performance.

CDNs are extensions of a GSLB system that is used to cache copies of static content (such as images and videos) in a distributed network of servers to bring the content closer to users. This shortens the distance data must travel and reduces latency.

CDNs can also optimize dynamic content delivery by using various optimization techniques such as multiplexing (where two or more data streams share a single connection to the host); SSL offloading (which removes SSL-based encryption from incoming traffic to unburden servers), dynamic site acceleration (which accelerates dynamic websites by quickly caching dynamic and unique content); and persistent connections (which direct all single-session user connections to the same backend server).

GSLB systems use traffic redirection protocols to transfer user requests to the most appropriate server. These solutions optimize the path traffic takes through the internet, avoiding congested and suboptimal routes and improving overall network performance.

Advanced solutions optimize connections and eliminate single points of failure by balancing loads from the beginning of a DNS request on a user’s device. Using real user monitoring (RUM) data and “last-mile visibility” (visibility into the final stretch of connectivity to the user device) these solutions can create more resilient networks and eliminate traffic bottlenecks and dependencies.

Advanced solutions can typically automate several different GSLB deployments, including:

Active-active. An active-active configuration comprises multiple active data centers (on-premises, private cloud, public cloud or hybrid cloud environments) where client requests are evenly distributed across centers, an ideal setup for global traffic distribution in a distributed environment.

In an active-active deployment, all sites are active, and services for a specific app or domain are connected to the same GSLB virtual server. Sites share data using the metrics exchange protocol (MEP), which includes information like load balancing and content-switching virtual server statuses, current connection numbers, packet rates and bandwidth usage.

When a user sends a DNS request, they’re directed to one of the active sites. If the request lands at a site, the GSLB virtual server there selects a load balancing or content-switching server and forwards its IP address to the DNS server, which relays it back to the client. The user then resends the request to the new virtual server at the provided IP address.

Active-passive. An active-passive setup comprises an active data center and a passive data center. In this configuration, passive sites are reserved exclusively for disaster recovery scenarios. In other words, passive data centers enter decision-making processes when all active sites are unavailable and a disaster event triggers a failover. If the active site experiences an outage, the passive site is activated.

When a user submits a DNS request to an active-passive GSLB configuration, the request may be directed to any site, but the system only deploys services from the active site (assuming it’s operational).

Global server load balancing is a dynamic technology that can automate and optimize network management for large enterprises operating in a range of industries. GSLB technologies also have myriad business applications, including:

Regardless of the sector or type of enterprise, deploying GSLB services can help organizations achieve:

If one server or data center goes down, GSLB can automatically reroute network traffic to the next best server to minimize (or even prevent) downtime and maintain a high-quality customer experience.

Advanced GSLB solutions can remove inline bottlenecks and single points of failure (SPoFs) from revenue-generating apps and websites by automatically steering traffic around deprecated or unavailable resources.

With GSLB solutions, teams can steer DNS traffic to the highest-performing, most available API endpoints to ensure the best app experience for users.

Many vendors offer SaaS options for GSLB services, helping businesses increase network flexibility and reduce costs by eliminating the need for physical and virtual appliances.

Using real user monitoring (RUM) data, GSLB programs can load balance connections all the way out to the end user, improving app response times and overall network performance.

GSLB services enable businesses to seamlessly integrate GSLB with existing load balancers and enhance their capabilities.

GSLB solutions can help organizations, especially those in highly regulated industries like healthcare and defense, stay ahead of government regulations by integrating country-specific regulations into local servers and GSLB forwarding preferences.

GSLB solutions make it easy to implement GSLB enterprise-wide with simple, self-serve onboarding and professional migration services to get teams up and running in no time.

Leading GSLB solutions can automatically scale resources up or down to accommodate fluctuations in network traffic.