Google officially gets in the battle for market share in the infrastrucuture-as-a-service (IaaS) area. What was only determined for a selected group of customers starting one year ago, the company from Mountain View has now made available for the general public as part of the Google I/O 2013. It’s about their cloud computing offering, Google Compute Engine (GCE).
News about the Google Compute Engine
With App Engine, BigQuery and Cloud Storage, Google has steadily expanded its cloud portfolio since 2008. What was missing was an infrastructure-as-a-service solution that can be used as needed to start virtual machines. The Google Compute Engine (GCE) released Google to its I/O 2012 in a closed beta, to use virtual machines (VM) with the Linux operating system on the Google infrastructure, which is also used by Gmail and other services.
Together with the Google I/O 2013, the GCE has now reached the general availability. Furthermore, Google has launched the Cloud Datastore, a by Google fully managed NoSQL database for non-relational data. Independent from the GCE the service provides automatic scalability, ACID transactions, and SQL-like queries and indexes. In addition, there is a limited preview of the PHP programming language for App Engine. With that Google wants to address developers and users of open source applications such as WordPress. Beyond that, the integration has been improved with other parts of the cloud platform such as Cloud SQL and Cloud Storage. Further, Google looks at the feedback of its users, that it should be possible to develop simple modularized applications on the App Engine. In response, it is now possible to partition applications into individual components. Each with its own scaling, deployment, versioning and performance setting.
Other major announcements include more granular billing, new instance types as well as an ISO 27001 certification:
- Granular billing: Each instance type is now billed per minute, where 10 minutes will be charged at least.
- New instance types: There are new micro and small instance types that are meant to process smaller workloads inexpensive and require little processing power.
- More space: The size of the „Persistent Disks“, which can be connected to a virtual instance have been extended up to 8.000 percent. This means that now a persistent disk can be attached with a size of up to 10 terabytes to a virtual machine within the Compute Engine.
- Advanced routing: The Compute Engine now supports based on Google’s own SDN (Software Defined Network) opportunities for software-defined routing. With that instances can act as gateways and VPN server. In addition it can be use to develop applications so that they run in the own local network and in the Google cloud.
- ISO 27001 certification: The Compute Engine, App Engine and Cloud Storage are fully certified with ISO 27001:2005.
Developer: Google vs. Amazon vs. Microsoft
First, the biggest announcement for the Google Compute Engine (GCE) is its general availability. In recent months, the GCE was held up by every news as THE Amazon killer, although it was still in a closed beta, and thus there was no comparison at eye level. The true time reckoning begins now.
Many promise from the GCE that Google creates a real competitor to Amazon Web Services. The fact is that the Google Compute Engine is an IaaS offering and Google due to its core business, have the expertise to build highly scalable infrastructures and to operate them highly available. The Google App Engine also shows that Google knows how to address developers, even if the market narrows here with increasingly attractive alternatives.
A lack of diversification
Having a look at the compute engine, we see instances, storage, and services for the storing and processing of structured and unstructured data (Big Query, SQL Cloud and Cloud Datastore). Whoever sees Google as THE Amazon killer from this point, should scale down its expectations once a little. Amazon has a very diversified portfolio of cloud services that enables to use the Amazon cloud infrastructure. Google needs to tie in with it, but this should not be too difficult, since many Google services are already available. A look at the services of Amazon AWS and the Google Cloud Platform is worthwhile for this reason.
Hybrid operation for applications
Google may not be underestimated in any case. On the contrary, from a first performance comparison between the Google and Amazon cloud, Google emerged as the winner. This lies inter alia in the technologies that Google is constantly improving, and on its global high-performance network. What is particularly striking, Google now offers the possibility to develop applications for a hybrid operation in the own data center and for the Google cloud. This is an unexpected step, since Google have been rather the motto „cloud only“. However, Google has been struggling lately with technical failures similar to Amazon, which does not contribute to the strengthening of trust in Google.
A potshot is the new pricing model. Instances are now charged per minute (at least 10 minutes of use). Amazon and Microsoft still charge their instances per hour. Whether the extension of the „Persistent Disks“ up to 10 terabytes will contribute a diversification we will see. Amazon is also under developers regarded as the pioneer among IaaS providers, which will make it not easier for Google to gain market share in this segment. In addition, Google may assume that, next to ordinary users, developers also do not want to play Google’s „service on / off“ games.
Amazon and Microsoft are already one step ahead
Where Google with its SaaS solution Google Apps massively tries to penetrate corporate customers for quite some time, the Compute Engine is aimed primarily at developers. Amazon and Microsoft have also begun in this customer segment, but long since begun to make their infrastructures respectively platforms attractive for enterprise customers. Here is still much work for Google, if this customer segment is to be developed, which is inevitably. However, in this area it is about much more than just technology, but about creating trust and to consider organizational issues (data protection, contracts, SLAs, etc.) as valuable.
Google’s problem: volatility
No doubt, Google is by far the most innovative company on our planet. But equally the most volatile and data hungriest. This also developers and especially companies both observed and should ask the question how future-proof the Google cloud portfolio is. If the compute engine is a success, don’t worry about it! But what if it is for Google(!) a non-seller. One remembers the Google Reader, whose user numbers were not sufficient enough for Google. In addition, the compute engine has another KPI, revenue! What does Google do when it’s no longer economic?