Google Enters the IaaS Club; Announces Google Compute Engine

As expected, Google has launched its Infrastructure as a Service offering called Google Compute Engine on the second day of Google I/O conference. Urs Hölzle, Google’s senior vice president of technical infrastructure announced that the Compute Engine is ready to tackle large-scale computing needs of the customers. The key differentiator of the platform seems to be the scale, performance and stability. Urs claimed that the internal network that connects the VM offers high performance. According to him, the service will offer 50 percent more computing per dollar than the competition.

Here are some of the early observations –

  • The platform will offer Linux VMs available in 1,2,4 and 8 cores with each core supporting 3.75 GB RAM.
  • Each compute unit is called GCU or GQ. 2.75 GQ represents the minimum power of one logical core.
  • Though it may be possible to host web applications, Google is emphasizing on batch processing, analytics and high performance computing.
  • The service will be eventually available across Americas, Europe and APAC.
  • Static IP address can be associated with each VM.
  • Persistent block storage devices can be attached to the VMs
  • There is a planned maintenance downtime for the VMs. Customers will be notified two weeks in advance. This may be present only during the developer preview.

We will bring the complete analysis of Google Compute Engine shortly!


Updates from around the world