Google Compute Engine

What is Google Compute Engine?

Google Compute Engine lets you run large-scale computing workloads on the same infrastructure that runs Google Search, Gmail, and Ads.

You can launch virtual machines on-demand, manage network connectivity using a simple but flexible networking solution, and access a verity of data storage alternatives from your virtual machines.

Google Compute Engine is in limited preview; please visit cloud.google.com to learn how to sign up.

Lots of virtual machines…

  • Highly scalable
  • High performance
  • Affordable

Features of Google Compute Engine


Google Compute Engine provides the following features:

1. Full Virtual Machines

Kernel-hosted virtual machines running either Ubuntu or CentOs. Launch 1, 2, 4, or 8 virtual core instances with 3.75GB of memory per virtual core.

2. Flexible Storage

Google Compute Engine offers three storage options to meet your varying needs:

i. Ephemeral disk

A simple block device that is bound to the lifecycle of the virtual machine (VM). When you stop the VM, the data is lost.

All data written to this device is encrypted to maintain strong data privacy.

ii. Persistent disk

A replicated network-connected storage service that is comparable to the latency and performance of local disks.

Data written to this device is replicated to multiple physical disks in a Google data center.

You can also create snapshots of your disks for backup/restore purposes and can mount these devices in a mode that allows multiple virtual machines to read from a single device.

As with our ephemeral disk, all data is encrypted at rest.

iii. Google Cloud Storage

Easily access your Google Cloud Storage data buckets from inside a virtual machine.

Our seamless authentication makes it easy to securely access your data without having to manage keys in your virtual machines.

3. Flexible Networking

We offer a network solution that makes it easy for you to connect your virtual machines to each other and to the Internet.

i. Isolation

Strong controls in our network stack protect our customers’ network traffic from unauthorized access.

ii. External IP addresses

Connect your virtual machine to the Internet with either static IP addresses or ephemeral addresses that are assigned to your machines.

iii. Configure FireWall

Control who can talk to which virtual machine(s) with a simple firewall system.

4. Open tooling

We offer a simple UI and command-line tool to configure and launch your virtual machines.

i. Open API

All of our tools are built on our open RESTful API, and we plan to release our tools as open-source to allow others to create customized tools for specific needs.

ii. Open ecosystem

We are working with providers such as RightScale, Puppet Labs, and OpsCode to deliver high-quality, familiar tools that work well in our environment.

Example of Google Compute Engine

Google Compute Engine

Guiding principles: How Google building Google Compute Engine


1. Strong Security

  • Network encapsulation
  • Disk encryption of data at rest

2. Consistency

  •  All about performance

– Processor allocations

– Storage latencies and throughput

– Network architecture

3. Open and Flexible

  •  Open API
  •  Open tooling
  •  Focus on the ecosystem

4. Proven 

  • One VM infrastructure

– For Google production workloads

– For your workloads

  •  Currently running Google production services

5. Powered by the Ecosystem

  •  Focus on building a vibrant ecosystem
  •  Strong core infrastructure
  • Our ecosystem partners

– Support mobility to the cloud

– Richer experiences

  • Think about services not servers

MapR’s Experience with Google Compute Engine


  1. Fast –  Virtualized public cloud rivals on-premise physical
  2. Easy –  Provision 1,000s of servers in minutes
  3. Cost effective –  Pay only for what you use

MapR Technologies

i. Open, enterprise-grade distribution for Hadoop

–  Easy, dependable and fast

–  Open-source with standards-based extensions

ii. Hadoop

–  Big data analytics

–  Hadoop inspired by MapReduce paper published by Google scientists Jeffrey Dean and Sanjay Ghemawat in 2004

iii. MapR is deployed at 1000’s of companies

iv. MapR Hadoop Cloud Service now available on Google Compute Engine


Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes:

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>