In a flurry of announcements made at their developers conference in San Francisco, Google today disclosed details of Compute Engine, an IaaS offering that runs Linux VMs on demand utilizing Google’s cloud infrastructure. Google Compute Engine (GCE) supports 1, 2, 4 and 8 virtual core VMs with 3.75GB RAM per virtual core. VM’s can be strung together to form clusters that can be managed using a network interface, and tools allow you to configure firewalls and control VMs using either a scriptable command-line tool or through a web interface. There’s also an API available that lets developers create their own management/control system. As is typical for Google, GCE will be made available on a limited preview basis, which you must sign up to receive. No date for general release has been set.
In related announcements the company also previewed the new AppEngine 1.7. AppEngine is considered restrictive by some because it forces you to code applications for specific APIs. As Cade Metz writes, “With raw virtual machines, developers can pretty much run whatever software they want, just as they can with Amazon EC2, the undisputed king of the cloud computing game.”
GCE offers sevearl storage options that allow you to either store data locally, utilize Google Drive or Google Cloud Storage. Asked about data security in the cloud, Google responded that:, "Google Cloud Storage benefits from the same proprietary security technology and procedures that protect Google data during billions of transactions daily. The system also supports OAuth 2.0 authentication."
Reactions at both the conference and over social media were mixed. @matthewsmall summed up the sentiment:
VMs, multiple types of storage, cluster networking, using Google's global network.
@thoughtspad wrote:
My reaction to Googl Comput Engin... Yay! My reaction when I saw pricing: Ouch! 1GB on Amazon's EBS $0.10/month, 1GB on Google $72/month... Google: you'v got to do a better job with pricing...
In a quick comparison, InfoQ found that Compute Engine’s pricing appeared higher. The base configuration for a single virtual core runs at $ 0.145 per hour compared to $.090 per hour for a similar configuration on Amazon’s EC2. According to the company, "Google Cloud Storage is available for a low monthly fee based on the amount of data you store. Storage and bandwidth usage are calculated in gigabytes (GB), where 1 GB = 2^30 bytes." These are not direct feature-for-feature comparisons, so you are encouraged to evaluate pricing based on your specific configuration. You may find Amazon's AWS Pricing Calculator useful in evaluating that side.
Few people have had an opportunity to kick the tires on GCE. In his review of GCE today, Joe Masters Emison writes:
Today, GCE's machine types are quite limited; all of them are built from even multiples of "1 CPU/2GB RAM" chunks. This is similar to SoftLayer, which has "1 CPU/1GB RAM" chunks, and Rackspace Cloud, which has both "1 CPU/1GB RAM" and "1 CPU/2GB RAM" chunks. Additionally, the amount of ephemeral disk space also goes up as the CPU/RAM goes up. The good news is that, unlike with the Rackspace Cloud, you can attach persistent storage to servers on GCE, so if you require a server with a lot of disk space but little CPU and RAM, you won't have to overpay. The bad news is that if you need a server with a lot of RAM but little CPU, or vice versa, you will overpay on GCE today.
Puppet Labs tweeted heavily about the Beta release of their new cloud provisioning module, PuppetForge, and their upcoming Puppet Enterprise + GCE. The GCE cloud provisioning Beta release is compatible with both Puppet open source and Puppet Enterprise.
Other solution partners include RightScale, OpsCode, Numerate, Cliqr and MapR.