The load balancer has been a core infrastructure component for high-performance Application delivery environments for decades providing resiliency, elasticity, and security. The name "load balancer" doesn't describe the full scope of functionality for this technology. In addition to load balancing traffic across servers, it also provides functions such as web app firewalls (WAF), global traffic management, analytics, observability, DNS, IP address management, and more.
Over the past decade, the application environment has dramatically changed, and this is creating a need for load balancers to evolve. The traditional load balancer is hardware based with integrated control and data planes. I want to be clear; these products have provided tremendous value over the past three decades, but the physical form factor is reaching the end of its life. One of the biggest problems with legacy load balancers is the lack of agility and scale. Hardware load balancers were designed for on-premises deployment in relatively static environments with static capacity. When multiple load balancers are deployed, and one is reaching capacity, it's impossible to move resources from one to another. Also, hardware load balancers do not fit in the software-defined virtual cloud operating model, and the multi-cloud world we live in today. This is just a limitation of hardware-based and legacy load balancers.
Today, the application environment has evolved into a highly dynamic ecosystem of public and private clouds, with edge computing coming fast. Also, containers, which are highly ephemeral in nature, are now mainstream. Containers can be spun up in a manner of seconds, and the hours it takes to spin up a VIP is far too slow, manual, and complex. The load balancer needs to evolve and be equally ephemeral in nature.
VMware’s NSX Advanced Load Balancer, which came to the company through the acquisition of Avi Networks, is designed with the future of application delivery in mind. The product is built for a higher level of agility with a cloud operating self-service model while streamlining operations, which leads to cost savings. This model, with a separate control plane and data plane, simplifies the management of load balancing and offers advanced analytics in these modern application environments.
Because the data plane can be deployed as individual auto-scale instances, each load balancing service can run independently of one another. When application traffic increases, the control plane can spin up new data plane (Service Engine) instances automatically with no manual configuration required. No tickets to deploy new applications or change the capacity are required. It is designed for automated scale-out with a self-service cloud operation model so IT can run at the speed of business. All of this can be done simply through a single control plane designed management interface.
This software-defined approach is a key in shifting to multi-cloud. While a business could use something like Azure Load Balancer in the Azure cloud, that service won’t work in AWS or GCP. Similarly, AWS Load Balancer won’t operate in Azure. On top of this, each cloud’s version of the load balancing function works differently, with different feature sets available. In a multi-cloud deployment, the business would need to configure each cloud load balancer independently, making consistency of services very difficult.
With NSX Advanced Load Balancer, the product can be deployed across multiple public and private clouds and managed through a single console creating a “fabric” of services that can be deployed, configured, and managed across the various clouds. It’s important to understand that compute is ever evolving, and multi-cloud is shifting to distributed cloud, which will bring in edge computing as well as private clouds. A single console also provides global visibility into the application performance and security across the different environments.
Digital businesses are successful only if they can move at speed. This is driving companies to modernize their compute and application environments, but the supporting infrastructure, such as load balancers, need to evolve as well. A cloud-native load balancer, such as NSX Advanced Load Balancer, ensures that all the critical services that optimize application performance are available at the speed the business requires.
Zeus Kerravala is the founder and principal analyst with ZK Research.
(Read his other Network Computing articles here.)