What is an Application Delivery Controller?

An Application Delivery Controller (ADC) is a load-balancing service that accepts TCP or UDP connections and maps each one to a healthy node in a pool of backend servers, using network address translation and a configurable algorithm. On Versa Operating System (VOS) devices, the same service distributes Analytics log traffic and manages redundancy across clusters.

Key takeaways

• An ADC load-balances connections across a pool of backend servers.
• On VOS, an ADC service defines a virtual IP and port pair to listen on.
• A default pool and backup pool give ADC services built-in failover.
• Controller-node ADCs distribute LEF log connections to Analytics clusters.
• VMS deployments use an ADC virtual IP to fail over between nodes.

On this page: Why it matters · How it works · Types · Application Delivery Controllers & Versa · FAQ

Why are Application Delivery Controllers important?

Any service with more than one backend instance needs something to decide which instance handles each connection – and to detect when one stops answering. That is the job of an ADC: a single point of contact, a virtual IP and port, in front of a pool of real servers, so clients never need to know how many exist or which is healthy.

Importance grows with redundancy requirements. Load-balancing a single site is only half the job; production deployments need instances – a server, a cluster, a data center – to fail without breaking client-facing connections. An ADC with a backup pool, not just a default one, turns local load balancing into a redundancy mechanism: connections keep flowing to a healthy backup automatically, no re-pointing needed.

What is an Application Delivery Controller?

An ADC sits between clients and a pool of backend nodes, presenting one address – a virtual IP and port – for clients to connect to. On each new connection, the ADC picks a pool member using a configured algorithm and rewrites the packet headers, through network address translation, so traffic reaches the chosen node and replies flow back through the ADC.

Health checking makes the abstraction reliable. The ADC continuously probes each pool member and removes unhealthy members from rotation without dropping sessions already established elsewhere. Many ADCs add a backup pool the service switches to if the entire default pool becomes unreachable, then optionally switches back once it recovers.

Types of Application Delivery Controllers

Three deployment models cover most of the category. Hardware appliances – dedicated devices such as F5 BIG-IP or Citrix (NetScaler) ADC – dominated the market for two decades: a physical box in front of a data-center server farm. Software and virtual ADCs deliver the same logic as a virtual machine, container, or embedded function, trading a dedicated appliance for elasticity and lower cost – the model most SD-WAN and cloud platforms use today. Global server load balancing (GSLB) operates one level up: instead of distributing connections among servers at one site, it steers clients – typically via DNS – toward the nearest of several separate sites, adding site-level failover.

Application Delivery Controllers & Versa

On a VOS device, the ADC is a built-in service, not a separate appliance: enabled per service node group and per tenant organization, then configured as an ADC service – a virtual IP and TCP/UDP port pair – backed by one or more ADC server pools. Each pool uses a configurable algorithm and an optional backup pool the service switches to if every default-pool server fails its monitor probe, falling back once the default pool recovers.

The most common deployment runs on Controller nodes, where the ADC load-balances log export (LEF) connections from VOS devices across an Analytics log collector cluster, toggling to a secondary cluster for active-backup or active-active redundancy. A second pattern fronts a Versa Messaging Service (VMS) high-availability pair: the ADC’s virtual IP gives every VOS device one stable address that always resolves to the active VMS node, even mid-failover.

Frequently asked questions

What is the difference between an ADC and a load balancer?

The terms are largely interchangeable. “Load balancer” describes the underlying function; “application delivery controller” is the broader marketing term for load balancers that also add health monitoring, network address translation, and sometimes Layer 7 content inspection – capabilities that go beyond simple round-robin distribution of connections.

How does failover work between an ADC’s default pool and backup pool?

The ADC continuously monitors the default pool with health probes. When every member of the default pool fails its probe, the ADC redirects new connections to the backup pool automatically; some implementations also support falling back to the default pool once it recovers.

What load-balancing algorithms do ADCs typically support?

Round robin distributes connections evenly in sequence; least connections and least response time favor the currently least-loaded node; source or destination IP hash keeps a given client pinned to the same node; weighted variants of each send more traffic to the more capable nodes in a pool.

Is Versa’s ADC the same as its Analytics log collector?

No. The ADC is a general-purpose load-balancing service on a VOS device; an Analytics log collector is a backend node it distributes connections to. On Controller nodes, the ADC’s most common job is spreading LEF log connections across a collector cluster.

What is Global Server Load Balancing (GSLB), and does Versa’s ADC provide it?

GSLB steers clients across geographically separate sites, typically via DNS. Versa’s ADC is a local load balancer: one virtual service distributes connections within a pool, with an optional backup pool for site-level failover – between primary and secondary Analytics clusters, for example – rather than DNS-based GSLB.

Related terms

If you’re reading about ADC, these are worth looking at next.