In case you missed it, last week Google made available load balancing capabilities to its Google Compute Engine offerings. In short, the new service is offering the following features:
- Load-balances ingress network TCP/UDP traffic over a specific set of Compute Engine virtual machines (VMs) within the same region
- Ensures that only healthy VMs are used to serve Internet requests through the use of HTTP-based health checks
- Easily handles spikes in load without pre-warming
- Configure the load balancer via command line interface (CLI) and a programmatic RESTful API
Presently, the service only provides layer three support but will perhaps have extended capabilities in the future as Google continues to build out its IaaS offering.
Presently, this load balancing service differs from something like AWS's Elastic Load Balancer service in that it can't support session affinity or SSL termination at the load balancer layer. However, Google Compute Engine's load balancer will allow you to load-balance UDP traffic if you have such a requirement, thus preventing you from having to manage an instance just to host a software load balancer.
The service is currently free until December 21st 2013.