Keeping Essential Public Services Always On: Multi–Data Center Failover with PAS-K GSLB

When Downtime Isn’t an Option

For a public institution delivering essential services to citizens, high availability is more than a technical metric — it’s a matter of public trust. An unexpected system failure or even a few minutes of downtime can do more than interrupt a single transaction; it can damage the institution’s reputation and, over time, push users toward alternative services. For the organization at the center of this case, continuous service availability — and the business continuity that depends on it — was a non-negotiable requirement.

Redundancy Alone Wasn’t Enough

To guard against outages, the institution had already built a hot-standby, active–standby architecture across two data centers: a primary site in Busan and a disaster recovery (DR) site in Seoul. Under normal conditions, the Busan site handles all user traffic; if it fails, that traffic is meant to move to the Seoul DR site so services stay online.

But standing up a dual data center is only half the problem. The harder question was how to move traffic between sites — automatically, instantly, and without users noticing. The team needed intelligent traffic control at the top of the infrastructure: a way to distribute load, continuously monitor the health of each site, and fail over the moment something went wrong. They also wanted to avoid bolting on a stack of separate, single-purpose appliances that would only add cost and operational complexity. What they needed was Global Server Load Balancing (GSLB) working hand in hand with their existing load balancing — delivered as efficiently as possible.

One Platform for Load Balancing and Failover

The institution deployed PAS-K, a PIOLINK Application Delivery Controller (ADC) that consolidates L4–L7 load balancing and GSLB onto a single platform. The same device performs Server Load Balancing (SLB) in front of application servers, Firewall Load Balancing (FWLB) in front of firewalls, and GSLB across the two data centers.

PAS-K’s GSLB engine continuously evaluates each site’s availability and network conditions, steering users to the optimal data center via DNS. During normal operations, it directs users to the Busan primary site. The moment a link failure or outage is detected in Busan, GSLB automatically fails traffic over to the Seoul DR site — with no manual intervention and no perceptible interruption for the end user.

Crucially, the PAS-K unit at the DR site also runs in active mode, performing the same health monitoring and traffic handling as the primary. This design eliminates any single point of failure between the two sites. And because PAS-K delivers SLB, FWLB, and GSLB natively — with no per-feature licensing to purchase — the institution consolidated what would otherwise require several separate appliances into one cost-efficient, easier-to-manage platform.

 

GSLB directs users to the Busan primary center and automatically fails over to the Seoul DR center, while a dedicated line keeps both sites in sync.

Near-Zero Downtime, Lower Complexity

The results were tangible. By pairing its dual data center architecture with PAS-K GSLB, the institution achieved near-zero-downtime failover between sites.

Service continuity improved: the Busan site carries traffic under normal conditions, and failover to the Seoul DR site is now fully automatic. Traffic management became genuinely intelligent — GSLB continuously monitors the availability of each site and updates its DNS responses the instant a failure is detected, keeping users on the fastest healthy path. Consolidating firewall load balancing and GSLB onto a single platform improved operational efficiency and removed the need for additional devices or a complex network redesign. And with integrated load balancing and no additional licensing, the institution lowered both deployment and operating costs — reducing total cost of ownership while leaving room to scale as demand grows.

“In the past, service outages were a major concern during system failures, but after adopting PAS-K GSLB, we’ve been able to maintain stable service operations. As a result, overall service reliability and quality have significantly improved.”

For public institutions — and any organization where downtime simply isn’t an option — this deployment points to a clear principle: true resilience doesn’t come from redundancy alone, but from intelligent, automatic failover. And when server load balancing, firewall load balancing, and GSLB are consolidated into a single Application Delivery Controller like PAS-K, that resilience arrives without the added cost and complexity of a multi-appliance architecture.