GCP: Google Cloud Platform

from the Hydra High Council Nov 6th 2018

What's Google Cloud Platform

Google Cloud Platform (GCP)is a Cloud Computing solution that offers IaaS, CaaS, PaaS, SaaS,FaaS.It'sbiggest competitors are Amazon Web Services and Microsoft Azure.

Google Cloud Platform Products

GCP is a suite of products and services that Googlehasbeen using internally with applications like Google Search, Google Maps, Gmail,andYouTube,and now are available tothegeneral public. Some of these products include:

App Engine-With App Engine,you can create highly scalable web applications and mobile backends. It's a fully manage service,so there is no infrastructure to manage or servers to maintain-you just focus on developing your application. App Engine willautoscaleyour application in response to how much traffic is received and is compatible with development tools, such asJenkins, Git, IntelliJ, Maven, PyCharm and Eclipse.

Compute Engine-With GCE you can create and run Linux, Windows or Customized versions of these operating systems. Optionally, you can build and run images of other operating systems. If you want to migrate your OS specific workload to the cloud and have full control of the resources to be assigned, then GCE will be the product to be used.

Kubernetes Engine-GKE is a fully managed cluster management and orchestration system for running your containerized applications. GKE is comprised of GCE instances which run Kubernetes. There is a master node whichmanages and interactswith a Cluster of Docker Containers using the Kubernetes API. Besides the master node, one or more nodes can existeach one running the Docker runtime and theKubeletagent.

Bigtable-Bigtableis a Petabyte-scale NoSQL database,ideal for applications that requirehigh throughput with very low latency. It's a great option for analytical and machinelearning applications, financial and IoT data, batch MapReduce operations and stream processing/analytics. Bigtable usesthe same open source API asHbase, so applications can be moved between Bigtable andHbase.

SQL-Cloud SQL offers fully managed relationaldatabases. You canrun MySQL and PostgreSQL databases on Cloud SQL without needing to do maintenance tasks,like applying patches, updates and managing backups. You can scale up by adding more CPU and memory or scale out by adding read replicas.

BigQuery-BigQueryis a fully managedanalytics data warehouse for storing and analyzing large datasets using SQL-like queries. Querying Big Data inBigQueryis easy and very cheap.

DataProc-DataProcis a fully managed service for running Hadoop/Spark clusters in a fast and easy way. Clusters are created very quick and it can scale using different machines types and number of nodes.

DataFlow-DataFlowis a managed service used to build and execute pipelines for data processing patterns like ETL, batch and streaming real-time analytics. Dataflow can also be used with other GCP services likeBigQuery,if you want to join yourBigQuerydata with other sources or with Cloud Pub/Sub for real-time streaming data.

ML Engine-ML Engineprovides machinelearning servicesso you can develop your own large-scale machine learningmodels or use one of the pre-trained models if they meet your requirements. You can use ML Engine for structured data like Recommendation, Fraud detection, forecasting or for unstructured data,like text, image and video analytics.

Why Google Cloud Platform

GCP provides better pricing than other cloud providers. On GCP you have per-second billing so you don't pay for minutes you don't use, giving you a true pay-only-for-what-you-use pricing model. You can create preemptible VMs that run at a much lower price that can be usedfor batch processing jobs. Also, GCP automatically applies a sustaineduse discount for instances running for more than 25% a month.

Traffic traversesthrough Google Private Global Fiber network,so you are talking over Google distributed Backbone, not the Internet.

GCP supports Live Migration of Virtual Machines. Google engineers can better address software and hardware issues without the need for you to worry about machine reboots.

GCP has global networks and regional subnets. Others Cloud providers offer regional networks and zonal subnets.

Google has continued to quickly expand the infrastructure for Google Cloud Platform. Currently, GCP has18 regions,55 zones,andover 100 points of presence across 35 countries with future regions in Zrich and Osaka.

Give it a try

Google Cloud Platform offers a free trial where you are provided witha$300 credit for 12 months,so you can try all GCP products. More info:https://cloud.google.com