Google Unleashes Their Public IaaS Cloud, Adds NoSQL Database
Google has opened the doors to its hotly anticipated Google Compute Engine and is now firmly engaged in the Infrastructure-as-a-Service (IaaS) battle with the likes of Amazon and Microsoft. This week’s Google I/O conference also introduced a new fully managed NoSQL database and the addition of PHP to the Google App Engine Platform-as-a-Service (PaaS).
Much like Microsoft did with their recently announced IaaS product, Google released the Google Compute Engine after an extended beta period. First introduced last June at Google I/O, the Google Compute Engine is now open to anyone who wants to deploy Linux virtual machines in Google’s robust data centers. Yesterday’s announcement highlighted pricing and storage features that differentiate Google Compute Engine from other IaaS competitors. A blog post from Google outlined the key points.
We’re now announcing several new capabilities that make it easier and more economical to use Compute Engine for a broader set of applications.
- Sub-Hour Billing: We heard feedback from our early users who wanted more granular billing increments so they could run short-lived workloads. Now all instances are charged for in one-minute increments with a ten-minute minimum, so you don’t pay for compute minutes that you don’t use.
- New shared-core instance types: Compute Engine’s new micro and small instance types are designed as a cost-effective option for running small workloads that don’t need a lot of CPU power, like development and test workloads.
- Larger Persistent Disks: We’re increasing the size of Persistent Disks that can be attached to instances by up to 8,000%. You can now attach up to 10 terabytes of persistent disk to a Compute Engine virtual machine, giving you plenty of persistent storage for a wide variety of applications.
- Advanced Routing Capabilities: Compute Engine now supports software-defined routing capabilities based on our broadSDN innovation. These capabilities are designed to handle your advanced network routing needs like configuring instances to function as gateways, configuring VPN servers and building applications that span your local network and Google’s cloud.
- ISO 27001 Certification: We’ve also completed ISO 27001:2005 certification for Compute Engine, App Engine, and Cloud Storage to demonstrate that these products meet the international standard for managing information security.
Unlike other IaaS providers that let you spin up virtual machines in any fashion, the Google Compute Engine requires every virtual machine to be part of a “project.” Google describes projects as “a totally compartmentalized world. Projects do not share resources, can have different owners and users, are billed separately, and managed separately.” Like AWS and other cloud providers, Google Compute Engine offers geographically distributed “regions” and “zones” where servers live. The currently available zones are in the United States and Europe, although the Network Pricing page indicates that Asia Pacific may be coming soon. The Google Compute Engine supports Debian and CentOS Linux instances with up to 52 GB of memory and 8 CPU cores.
Google also announced the preview of a new database offering.
Google Cloud Datastore is a fully managed and schemaless solution for storing non-relational data. Based on the popular App Engine High Replication Datastore, Cloud Datastore is a standalone service that features automatic scalability and high availability while still providing powerful capabilities such as ACID transactions, SQL-like queries, indexes and more.
A fully managed cloud database means automatic replication, automatic scaling to meet usage needs, active monitoring by Google staff, and no downtime caused by maintenance windows. Already with libraries for Node.js, Python, and Java, the Google Cloud Datastore is targeted at developers who want to store “entities” in a high-performing schema-less database. While it does support RDBMS-like features such as transactions and entity relationships, it doesn’t support capabilities like joins or subqueries.
The addition of the PHP runtime to the Google App Engine – and the ability to partition applications into separate scaling groups – was the final major cloud announcement from Google at this week’s conference. A Google blog post described this update to their PaaS platform.
App Engine 1.8.0 is now available and includes a Limited Preview of the PHP runtime - your top requested feature. We’re bringing one of the most popular web programming languages to App Engine so that you can run open source apps like Wordpress. It also offers deep integration with other parts of Cloud Platform including Google Cloud SQL and Cloud Storage.
We’ve also heard that we need to make building modularized applications on App Engine easier. We are introducing the ability to partition apps into components with separate scaling, deployments, versioning and performance settings.
Up until now, the Google App Engine only supported applications written in Java, Python, and Go. With the addition of PHP, Google has included the language of choice for hundreds of millions of Internet websites. While trendy options like Node.js or Ruby on Rails may have been a more celebrated choice, PHP support was the most commonly requested feature for the platform. Wired sees some strategy at play in this decision.
It makes sense: even companies and developers who are dumping PHP still have old apps that they need to run, and they want to run all their apps in one place.
Now, thanks to Google, it will be that much easier to keep that PHP code kicking around for a few more years.
Uwe Zdun, Rafael Capilla, Huy Tran, Olaf Zimmermann Mar 09, 2014
Olav Maassen, Liz Keogh & Chris Matts Mar 08, 2014