August 07, 2013, 1:55 PM — Google Wednesday rolled out load balancing features to its public cloud service, allowing customers to automatically scale up and down virtual machines to accommodate unexpected spikes in demand.
The rollout comes just a few months after Microsoft improved its Azure cloud service with new auto-scaling features. Both companies are effectively playing catch-up with leading IaaS provider Amazon Web Services, which already offers such features.
[WHEN DISASTER STRIKES:World Savings Robots]
Load balancing is a "critical" feature for any highly scalable cloud deployment, Google engineers wrote in announcing the company's service today. It allows Google Compute Engine (GCE) to automatically and intelligently route traffic across a collection of servers. This replaces a manual process where new virtual machines (VM) would be provisioned by the user. It's intelligent because the system can automatically check to ensure the VMs are healthy and can accept traffic. The load balancing can be configured either using a command line interface or through APIs.
Earlier this summer Microsoft released new auto-scaling features to its Azure cloud platform. That allowed similar functionality, with the ability to scale compute resources up or down compute resources in its cloud.
Both auto-scaling and load balancing are typically used to handle traffic spikes or large-scale increases or decreases in resources. The difference between the two is that auto-scaling spreads CPU workloads across a predefined cluster of servers, whereas load balancing allows network bandwidth traffic to be distributed across clusters of servers on the fly, according to a forum post on AWS's website.