The server age is over, the cloud age is here – this is how we could characterize our feelings from our visit to the Google Cloud OnBoard conference held last week at the Congress Centre in Prague. Google presented interesting facts and news about the Google Cloud platform and many modules enabling efficient work with data on the cloud. In the following lines we bring you a report.
The Google Cloud OnBoard conference was packed with content and information. The basic capabilities offered by Google Cloud Platform and its individual parts were presented. The entire cloud infrastructure is evolving at a relatively fast pace and responds flexibly to customer requests for new functionality, as evidenced by the recent implementation of the TensorFlow machine-learning framework.
There was no shortage of opportunities to run virtual servers on the platform and options for storing different types of data. Here, Google has covered a wide range of use cases, such as block or object storage, which can be easily connected to a global CDN to ensure fast distribution of content around the world.
Other types of storage introduced include Cloud SQL, which can provide compatibility with applications requiring MySQL or, more recently, PostgreSQL, which can ensure high data availability even if one of the database servers becomes unavailable. Another supported storage solution is Cloud Bigtable, used for storing large amounts of data.
After lunch, the conference program continued with an introduction to the possibilities of using container virtualization in Google Cloud, especially the Kubernetes orchestration platform, which we use and Google provides preconfigured clusters built on this platform within its cloud. Another new feature is the Google App Engine, a platform used to run applications in the cloud, unfortunately here Google has not yet heard the call from developers in Ruby on Rails and Ruby is not yet an officially supported language, although there are community (or rather partner) templates that run Ruby on Rails on the platform.
We also got an overview of how to develop, deploy, and monitor applications on Google Cloud Platform. And another very interesting area is the possibilities for big data solutions and machine learning. Perhaps everyone was very surprised by the demonstration of BigQuery’s big data capabilities, where a query was run over the entire contents of the Wikipedia encyclopedia (327 GB of data – approx. 70 DVDs) to find all articles containing a mention of Prague. The query took about 3 seconds, on a normal database server it would have taken at least several minutes, so it is a speedup of several orders of magnitude. This technology is commonly used to examine human DNA. The reference company for the demonstration was a company that does this processing and works on data of 10 PB (about 10,000 hard drives of a typical office PC).Finally, there were the possibilities of cloud artificial intelligence and a new development – Cloud Video Intelligence, a service that is able to mine information from video. Until recently, Google provided this service under the name Vision API for images only – in a hands-on demonstration, we saw a picture of a taproom and a person drinking a beer. The system is able to recognize the person’s mood and other data such as gender, objects in the image and other relevant information.
The conference allowed us to learn about the capabilities that Google Cloud Platform offers, however we are familiar with and use some of the technologies presented. If you are interested in clod solutions, need specific advice or have an assignment in this area, please contact us, we keep up with the trends.