Network Load Balancer Now Generally Available on IBM Cloud VPC

3 min read

We are excited to announce the release of the Network Load Balancer (NLB) offering.

This announcement extends the Load Balancer portfolio on IBM Cloud VPC Gen 2. Customers now have the option of using either an Application Load Balancer or a Network Load Balancer to distribute application and/or network traffic to Virtual Server Instances or Kubernetes clusters.

IBM Cloud Virtual Private Cloud (VPC) provides a logically isolated environment on IBM Cloud for your mission critical applications. To increase the reliability of your application and avoid any disruption to your end users, application or network load balancing capabilities are recommended to distribute incoming user traffic for better performance, lower latency, and higher availability.

Introducing Network Load Balancer on VPC

IBM Cloud offers two load balancers on VPC: Application Load Balancer and Network Load Balancer. 

Application Load Balancer operates at Layers 4 and 7 of the OSI Model and performs the functions of classic load balancers by distributing user requests across multiple targets.

However, there are occasions where you may prefer a single static IP for the load balancer rather than a Fully Qualified Domain Name (FQDN) or you would like to avoid capability restraints because of the cached DNS mapping till Time-to-Live (TTL). 

Network Load Balancer works in Layer 4 and is best fit for business-critical and interactive workloads that require high throughput and low latency.

IBM Cloud offers two load balancers on VPC: Application Load Balancer and Network Load Balancer. 

Why is a Network Load Balancer needed?

You may be asking why a separate Network Load Balancer is needed if the Application Load Balancer supports Layer 4. 

Often, a client may submit a request that is fairly small in size, with little performance impact on the load balancer; however, the information returned from the backend targets (VSI or container workloads) can be significant—perhaps several times larger than the client request. 

With Direct Server Return (DSR), the information processed by the backend targets is sent directly back to the client, thus minimizing latency and optimizing throughput performance.

What is Direct Server Return? Why is it important?

One of the key benefits of using NLB is the incredible performance of the load balancer itself. Its ability to process numerous, simultaneous requests and quickly return the results in an efficient manner is what makes Direct Server Return stand out from other load balancer options. Let’s walk through an example.

In the diagram below:

  1. A client initiates a request that is received by the load balancer (see purple lines). 
  2. The load balancer will send that request to the target—either VSI or Kubernetes cluster (see blue lines).
  3. With an Application Load Balancer, the target will process the request and send the results back to the client through the load balancer (see black lines).
  4. With Network Load Balancer, the target will process the request and send the results directly to the client, bypassing the Load Balancer (see green arrows) resulting in greater overall performance (i.e., larger data throughput and lower latency).    
What is Direct Server Return? Why is it important?

Benefits of Network Load Balancer

Because of the tremendous ingress and egress network performance capabilities of Network Load Balancer, it is ideal for audio/video or gaming workloads. Additional benefits include the following: 

  • Optimized load balancer performance: The Network Load Balancer is only required to process incoming traffic, which allows it to quickly distribute requests and to support higher traffic than the Application Load Balancer. 
  • Clients receive results faster: With the Direct Server Return feature in Network Load Balancer, target server results are sent directly to the client, bypassing the load balancer and resulting in an optimized, faster way to deliver results to the client.   
  • Faster failover recovery: The Network Load Balancer uses single, highly available Virtual IP (VIP) that can be used directly instead of using an assigned Fully Qualified Domain Name (FQDN). The VIP option provides faster recovery in the event of failures compared to a DNS-based availability that the Application Load Balancer uses.

Learn more about Network and Application Load Balancers for IBM Cloud VPC

Be the first to hear about news, product updates, and innovation from IBM Cloud