Podcast: Play in new window | Download
Computational load is the amount of demand that is being placed on a computer system. “Load” can take the form of memory, CPU, network bandwidth, disk space, and other finite resources.
When we design systems, we need to prepare for high-load events. On a social network, people are much more active in the mornings. On an e-commerce site, Black Friday causes many more users to come online for discount shopping. Our distributed application must be able to scale in response to these spikes in traffic.
Cloud computing has changed the popular software architecture patterns, and load balancing has changed along with it. With on-demand, infinite infrastructure, we don’t need to worry about ordering servers and provisioning. With infrastructure as code, it becomes simpler to manage lots of deployable units–so we can break up our monolith into microservices, and have hundreds or thousands of virtual machines or containers running.
Enterprises that were started before cloud computing have large on-premise server deployments–but today, many of them also use the cloud. The cloud can be used to augment their classic on-prem deployments with cloud platform-as-a-service features. The cloud can also be used as a reliable way to scale during high load events.
Today, a common architectural pattern is to have your application broken up into services. Each of those services has multiple instances. When the load on a particular service is under lots of demand, you create more instances to handle the increased load. How do you monitor the load on each service? How do you know when to spin up new instances of the service?
Load analysis and load balancing across different services can be implemented by placing “agents” throughout your infrastructure. These agents gather data about services and service instances, and route that data to a centralized place. The centralized “control plane” can be used to make decisions about load-balancing and traffic routing.
Ranga Rajagopalan worked on networking at Cisco for a decade before co-founding Avi Networks as CTO. Avi Networks builds modern load balancing software, and in today’s episode, Ranga describes the requirements of load balancing. We talked about the evolution of network infrastructure, the impact of the cloud, and the technical decisions that his team has made when architecting Avi Networks. Full disclosure: Avi Networks is a sponsor of Software Engineering Daily.
Transcript provided by We Edit Podcasts. Software Engineering Daily listeners can go to weeditpodcasts.com/sed to get 20% off the first two months of audio editing and transcription services. Thanks to We Edit Podcasts for partnering with SE Daily. Please click here to view this show’s transcript.
Pingback: Edition 17# – Late Night Link()