by: Senior Director of Network and Multi-Cloud Services Swapnil S. Hendre and Solution Services Manager Preethy G
The load balancer landscape is changing rapidly as more workloads continue to be transitioned from monolithic to modern applications. There is also growing demand for load balancers to be elastic. Our legacy load balancer lacked flexibility, cloud capabilities and elasticity. Additionally, native load balancers provided by major cloud providers were not advanced enough to support our application requirements.
The VMware NSX® Advanced Load Balancer™ platform delivers automated application services, including load balancing, application analytics, predictive autoscaling and security. The platform, built on software-defined principles, runs on standard x86 servers, virtual machines (VMs) or containers, which aligns with our automation and self-service goals. The controller runs in the control plane and manages service engines (SEs) that run in the data plane. This separation enhances our availability and elasticity and provides unified management. See Figure 1.
Figure 1: VMware NSX Advanced Load Balancer key features.
Balanced advantages
Some key benefits we observed by implementing NSX Advanced Load Balancer include:
Analytics
o NSX Advanced Load Balancer provides detailed real-time logs that help us quickly isolate the root cause of service disruption. This increased visibility for not just administrators but also for application teams to view and troubleshoot issues.
o The “health scores” which are calculated based on metrics like security, unavailable pool members, number of client hits missed (packet retransmission), and speed of response (latency) are beneficial for understanding the high-level view of our environment and quickly pinpoint affected services.
Elasticity
o Our legacy load balancers lacked elasticity, and we had to procure physical load balancers in anticipation of future growth. This frequently led to unused capacity.
o Due to the 100 percent virtual and elastic nature of NSX Advanced Load Balancer, we directly integrated the platform with the underlying infrastructure (VMware vCenter®, AWS, VMware Tanzu® Kubernetes Grid™) to enable horizontal scaling, with the SEs created and deleted based on number of virtual services. This significantly reduced costs for load balancer infrastructure.
Automation
o Day-to-day operations, such as creation of virtual services, pools, SE creation, and IP and DNS allocation are now automated, thereby significantly reducing human error and engineer-hours required.
Use cases
The NSX Advanced Load Balancer proved its mettle in numerous uses cases. Here are just a few examples.
Use case 1: remote site migration
In the beginning, we started simple—by replacing our legacy load balancers in remote sites. The approach we took for this migration was to pre-stage the configuration in NSX Advanced Load Balancer and perform an automated cutover. Scripts were used to disable the address resolution protocol (ARP) of the virtual services in the legacy load balancer and enable them in NSX Advanced Load Balancer, thereby also providing an option to rollback if needed. In this way, we achieved a seamless migration, with a service disruption of less than 10 seconds overall. See Figure 2.
Figure 2: Logical architecture of remote site migration.
Use case 2: application delivery
VMware is in the process of migrating monolithic applications to Kubernetes, and we use Tanzu Kubernetes Grid to manage Kubernetes container environments. While Kubernetes simplified our app delivery, a major missing piece was eliminating manual effort while configuring the virtual IPs (VIPs). On average, after application deployment was completed, virtual service configuration used to take three days, including submitting a formal request, manual configuration and troubleshooting due to misinterpreted requirements. As you can imagine, this was a major speed bump in our application deployment. Our goal was to eliminate all manual steps involved and enable self-service load balancer configuration and domain name system (DNS) registration.
We achieved our goal by deploying the NSX Advanced Load Balancer integrated with the Tanzu Kubernetes Grid via the Avi Kubernetes operator (AKO). The AKO is deployed directly in the Kubernetes pods—the smallest, most basic deployable objects in Kubernetes—and ensures that all the namespaces, ingresses and services created in Tanzu Kubernetes Grid are automatically created in the load balancer. Once the integration was complete, the process became completely automated. The application team only had to kick off the YAML script, which automatically creates PODs, Kubernetes services, deploys service engines, creates ingress pools, VIPs, health monitors, as well as complete DNS registration—all in 30 seconds. See Figure 3.
Figure 3: NSX Advanced Load Balancer integration with Tanzu Kubernetes Grid reduced application delivery from three days to 30 seconds.
Use case 3: VMware Cloud on AWS deploymentent
VMware is in the process of migrating critical workloads to the public cloud to leverage advantages like agility, elasticity, resiliency, flexibility, and cost savings. The NSX Advanced Load Balancer is integrated, out of the box, with our public cloud deployments. Key use cases include:
- VMware Horizon® virtual desktop infrastructure (VDI) servers in two VMware Cloud™ on AWS software-defined data centers (SDDCs)
- Native AWS deployment of mission-critical workloads using the Tanzu Kubernetes Grid platform to increase application service-level agreements (SLAs) to 99.99 percent. See Figure 4.
Figure 4: Public cloud deployment (VMware Horizon 7 on VMware Cloud on AWS—SDDC1 + SDDC2).
Use case 4: data center migration
In hindsight, this was our most complex use case—migration of high-value applications from a legacy load balancer to the new platform. Rather than taking a big-bang approach of migrating the entire data center in one shot, we started by grouping virtual services by applications and performed migration on a per-application basis. See Figure 5.
Figure 5. Logical architecture of data center migration.
What we learned continues to impress
A few years have passed since deployment, and we’ve learned a lot, including how load balancing can work effectively in the cloud era. Outage scenario mitigation is easier as there is support for SE horizontal scaling, decreasing downtime. The analytics capabilities make issue isolation a simple affair while reducing the burden of operations and application teams by increasing visibility into various parameters. Load balancer is not a black box anymore.
There is much more to talk about, so contact your account team to schedule a briefing with a VMware IT expert. No sales pitch, no marketing. Just straightforward peer conversations revolving around your company’s unique requirements.
Also check out our case study on this topic: How VMware IT Achieved Perfect Balance with VMware NSX Advanced Load Balancer.
VMware on VMware blogs are written by IT subject matter experts sharing stories about our digital transformation using VMware products and services in a global production environment. To learn more about how VMware IT uses VMware products and technology to solve critical challenges, visit our microsite, read our blogs and IT Performance Annual Report and follow us on SoundCloud, Twitter, and YouTube. All VMware trademarks and registered marks (including logos and icons) referenced in the document remain the property of VMware.
