In the networking world load balancers by definition are all about balancing the traffic load to ensure availability and service performance. IBM is now looking to upend traditional approaches to load balancing with a new service called NS1 Connect Global Server Load Balancer (GSLB).

The new service is an expansion of IBM's NS1 domain name system (DNS) and traffic management capabilities. IBM acquired managed DNS provider NS1 last February and expanded the offering with the NS1 Connect service in September. With the new NS1 Connect GSLB service IBM is aiming to use the power of DNS and real-user monitoring (RUM) data to disrupt the market and bring next-generation load balancing capabilities to organizations across the globe.

The NS1 Connect GSLB service is being positioned by IBM as an alternative to current in-line GSLB solutions offered by various vendors. According to IBM, routing traffic through in-line load balancers introduces several downsides including a lack of visibility into last-mile traffic. If the in-line load balancer goes down, so does the application. Adding more load balancers helps with resilience but does not fully solve the issue. IBM believes its DNS-based approach will have a strong impact on the load balancing market and change views on traditional in-line solutions.

“Our customers have been using DNS for many years to steer traffic between back-end resources like clouds and CDNs,” David Coffey, VP, product management and NS1 chief product officer, told SDxCentral.  “IBM NS1 Connect GSLB expands and extends this concept more broadly to fill the functionality and visibility gaps of today’s in-line load balancing technologies.”

At the core of the NS1 business is the capability to optimally route DNS queries to help internet traffic get to the right destination in a consistent, reliable and fast approach.

Bringing that same model to the GSLB world has some advantages, Coffey said. He explained that traditional GSLB solutions are combined with in-line load balancers. These all have the same architecture – they sit between end user devices and back-end resources. Coffey noted that this naturally creates choke points and single points of failure while ignoring the global connectivity performance aspect.

“IBM’s GSLB solution balances loads at the DNS level, so it's not directly in the application traffic flow, thus avoiding potential issues that would compromise performance or availability,” Coffey said.

While DNS is helpful in optimizing load balancer operations, there is still a real need to actually understand what the user experience is for a specific type of network traffic. That's where Real User Monitoring (RUM) fits in.

Coffey said that RUM has been a core element of the NS1 platform for several years and was originally part of the NS1 Pulsar feature. The way RUM works with NS1 Connect is that organizations configure their web applications with JavaScript tags that measure key application performance indicators like latency and availability whenever someone uses their application.

“This vast amount of RUM data is collected and processed by NS1 at a global level and represents a statistically significant sample of internet performance,” he said.

According to Coffey,  RUM data gives organizations a real-time picture of system availability and latency across all providers, helping them compare response times of different back-end resources based on the end user’s location and service provider. This allows them to make informed decisions based on real world conditions.

Beyond the RUM and DNS capabilities of NS1 Connect GSLB, Coffey also emphasized that the IBM approach is also different from a traditional load balancer approach in terms of how it is priced.

“IBM’s GSLB solution is a SaaS offering, so the cost of physical or virtual devices, licenses, and maintenance is eliminated,” he said.

Coffey added that the offering is based on a price per query model, so organizations pay for what they actually use vs. the traditional capacity-based pricing of in-line GSLB solutions.

“The savings will vary by in-line vendor and deployment types, but we believe that the savings will be significant - millions of dollars every year for our largest customers,” he said.