Google Cloud Platform

In April 2008, the Google developer team announced a closed developer preview of their new Platform-as-a-Service offering: Google App Engine. Google invited 10,000 lucky (and brave) developers were to test and provide feedback on an early version of App Engine. By May, that number had increased to 75,000 active developers; Google announced fully open signups, making App Engine available to the masses.

In the years that followed, Google released a steady stream of products and features. With services such as Google Cloud Storage in 2010, Compute Engine in 2013, Cloud SQL in 2014, and Kubernetes Engine in 2015, Google has built out a diverse and comprehensive suite for developing cloud-native solutions. During this time, Google looked to expand their domain into varying areas such as infrastructure management, data analytics, Internet of Things, and machine learning. By 2017, Google had established data centers in 39 zones across 13 regions.

With fierce competition among the major public cloud providers, Google is looking to establish itself as a market leader. With services such as BigQuery, Bigtable, Cloud Pub/Sub, and Dataflow, Google has thrown down the gauntlet in the data analytics arena. With a robust global infrastructure and experience running applications at scale, Google is looking to win over developers wanting to build solutions that support small groups of early adopters and effortlessly scale to support floods of users as applications go viral. With decades of experience providing highly available web services such as Search and Gmail, Google is positioned to redefine reliability in the cloud.

Today, the Google Cloud Platform catalog includes several products and services that cover a large number of use cases and industries. Core services such as Compute Engine and Cloud Storage enable teams to build virtually any solution, while many specialized services such as the Cloud Vision API greatly lower the barrier of entry for teams to tackle more specific problem spaces. As Google moves full steam ahead into the public cloud space, the number of both core and specialized products and services continues to grow at breakneck speed, as shown in the following graphic:

The Google Cloud Platform catalog contains many products, covering a wide array of use cases

Standing on the shoulders of giants

Google Cloud Platform is the product of decades-long experience running some of the largest and most successful web services in history. The infrastructure Google offers in GCP is the same infrastructure Google uses internally, meaning customers directly benefit from the wealth of hard-won knowledge and ingenuity Google has amassed through running many of their well-known large-scale services. Extreme reliability and security are established norms at Google, and these qualities are deeply ingrained into GCP's underlying infrastructure.

Google also embeds and applies this knowledge and experience to their managed services. Google App Engine is the direct product of Google's expertise managing web-scale services and is designed to make scalability a non-issue. With easy-to-use service integrations and managed autoscaling, engineers can develop against simple interfaces to quickly create web services that scale to any load. Likewise, Kubernetes (and by extension Google Kubernetes Engine) is the result of Google's experience, successfully orchestrating massive numbers of web services via the internal data center scheduling and orchestration platform known as Borg. BigQuery is the result of externalizing Google's own analytics platform, called Dremel. Google Bigtable is built on top of Google's powerful internal lock system, Chubby. Cloud Datastore builds on Bigtable clusters to provide easy-to-use managed document stores. Cloud Storage, BigQuery, and Bigtable are all built on top of Google's large-scale clustered filesystem Colossus (originally Google File System (GFS)). The point is, when you use GCP, you are the direct beneficiary of Google's success.

A world-class global presence

Google's 13 regions are connected by the first multi-tier global fiber network from a major public cloud provider. With over 100 points of presence, Google Cloud offers your users low latency no matter where they are in the world. This private fiber optic network is the backbone of Google's own global presence, made available to GCP customers. On top of this, Google offers powerful networking tools for easily building out your own network architecture. These tools include fully software-defined networks, self-adjusting network routing between on-premises networks and the cloud via Cloud Routers and VPNs, and dedicated interconnection to bring Google's stellar network to your door.

For a clear visual, please refer to the image of Map of regions and fiber network mentioned at https://cloud.google.com/about/locations/.

Building globally available services comes with a distinct set of problems, which Google is committed to addressing. For instance, as a user base grows, geographical issues such as data consistency become more challenging. To solve this problem, Google created Cloud Spanner—a strongly consistent relation database that scales to thousands of nodes across the world. Content-heavy service providers look to provide a consistent experience across their user base. On a global scale, this can become challenging due to network limitations such as latency and congestion. To address issues like these, Google offers worldwide CDN services via multi-regional Cloud Storage buckets. To enhance your global reach further, Google offers a range of extremely powerful load balancing solutions. With features such as anycast IP for simplified DNS, health check integrations, and content-aware routing, Google's load balancers make it easy to reap the benefits of a global presence.

Choosing your own adventure

Google's service offerings give developers the freedom to choose how much control they want over the system. For example, a team looking to build a data analytics process can choose from solutions ranging from fully managed (Dataflow), partially managed (Dataproc), to fully self-managed (Hadoop on Compute Engine). On the application side, solutions can range from a fully serverless model with Cloud Functions, managed PaaS solutions leveraging App Engine, the partially-managed Google Kubernetes Engine, to the extreme of running applications on Compute Engine with load balancers, managed instance groups, and backend services.

This continuum of service offerings is common across many areas of Google Cloud and embodies the philosophy of developer and operations enablement. The decision on which solution best fits a specific need is, of course, not entirely so clear cut, but it is worth noting that the services offered on GCP are as diverse within specific problem spaces as they are across separate problem spaces. Google looks to provide specialized tools rather than adopt a one-size-fits-all approach.

Leading the way for big data

Google is betting big on data. With so much business value being driven by data analytics, many modern technology companies are betting big on big data as well. Google offers a cohesive suite of tools to help you quickly and easily build out analytics solutions without getting bogged down in infrastructure management. From world-class data warehousing and analytics with BigQuery, to self-balancing data-processing pipelines on Dataflow, Google Cloud has tools to fit any need.

Teams can quickly start their data migration journey by moving existing Hadoop and Spark workloads to managed clusters on Dataproc. Rounding out these tools are services such as Pub/Sub messaging, Dataprep, and Google Data Studio for a fully managed, serverless, democratized analytics platform.

To further drive predictive analytics, Google is dedicated to bringing machine learning to the masses. With Cloud Machine Learning, users can easily get started with the powerful Google-born open source TensorFlow framework. This means developers can leverage the same tools Google uses internally to accomplish tasks such as speech and image recognition, all the while maintaining deep integrations with the rest of the big data offerings on GCP.

The Open Cloud and innovation

Google is making waves and building a reputation as the Open Cloud. Building on the core belief that developers should want to use GCP, Google consistently adopts and drives open standards and open source tools and frameworks. By open sourcing projects such as Kubernetes and TensorFlow, these projects are able to grow rapidly and organically. Instead of creating vendor lock-in, Google is then able to capitalize on these open source projects by providing the best developer experience on top of them, as seen in Kubernetes Engine and Cloud Machine Learning.

By adopting and adhering to open standards, Google further reduces the risk of vendor lock-in, and provides a lower barrier to entry for teams looking to move to managed services. This can be seen in a number of products, such as Cloud Bigtable, which adheres to the open-source Apache HBase interface, and Cloud Endpoints, which adheres to the OpenAPI specification. By working together with the wider community, Google creates a transparent, symbiotic relationship with developers that facilitates progress throughout the technology industry.

In addition to driving open sourcing and open standards, Google Cloud continuously innovates on ways to make more solutions feasible for organizations of all sizes. By providing per-second billing on compute resources, more teams can afford to build out massive-scale solutions such as spinning up hundreds of virtual machines for short-lived but intensive workloads. Innovating on the traditional approach of provisioning virtual machines, Google offers custom machine types that help developers optimize their use of cloud resources. With very competitive pricing, automatically applied sustained-usage discounts, proactive alerting on underutilized resources, and generous free tiers, Google helps teams minimize costs. Very often, Google Cloud is not just the best choice; it's the cheapest.

Dedication to customer success

The folks at Google understand the perceived risks in adopting the public cloud. Giving up control over your infrastructure can be scary. Every business is unique in their technology needs, and there are many unknowns. Instead of a one-size-fits-all model of cold documentation and endless FAQs, Google is dedicated to providing a customer-centric experience to help you build the best possible solutions on GCP. The Google Cloud team has internalized this ideology and formalized it into the practice of Customer Reliability Engineering.

With Customer Reliability Engineering, or CRE, Google is taking a vested interest in the reliability of your applications. This goes beyond the reliability of the underlying cloud services your application is running on. CRE realizes that the primary concern of teams running applications on Google Cloud is not the reliability of GCP itself, but rather the reliability of the applications those teams are responsible for. The reliability of the Google Cloud infrastructure is, of course, a factor in the reliability of your applications (and those grounds are well covered—see https://landing.google.com/sre), but Google is determined to go beyond delivering a stable platform to ensuring that the applications running on the platform are built for reliability as well.

Bottom-up security

A major point of contention for some businesses considering migrating to a public cloud is security. Customers trust you with safeguarding their identity and privacy—a responsibility that should be held in the highest regard. Google understands the weight of this responsibility, and the engineers of Google Cloud are dedicated to extending the same level of security to your customers that they provide to their own. From purpose-built security chips on GCP servers to globally available private network solutions, Google is dedicated to providing security at all levels of the platform.

Google also understands the importance of making security easy. With design features such as encryption at rest and services such as the Data Loss Prevention API and Cloud Key Management, the Google Cloud team is driving customer security by making it accessible and approachable. In bringing security to the forefront of their offerings, Google is helping to make security one of the primary motivations for public cloud adoption.

In good company

Since you are reading this book, it is assumed that your team is either considering leveraging Google Cloud or is already doing so. In doing so, you will be joining a group of diverse and rapidly growing companies across business sectors and geographic locations. From large enterprise companies such as The Home Depot and Coca-Cola, to technology companies such as Evernote and Vimeo, many people are finding that Google Cloud Platform has the tools and services they need to succeed.

This growing traction also creates enormous opportunity for businesses to learn from each other in the wide range of problems being solved on GCP. For example, looking at how Spotify leverages Google Cloud to stream songs to their customers, we can learn about the viability of multi-regional Cloud Storage buckets as a global CDN. By studying Niantic, we can see the power of Google Kubernetes Engine to rapidly scale applications to thousands of nodes. Read about these companies and many more at https://cloud.google.com/customers.

In addition to the many amazing companies already leveraging Google Cloud, Google is strongly focused on developing a network of partners with major companies such as Cisco, Pivotal, and Salesforce. These partner companies are offering services and integrations that make it easier than ever to bring your business to the cloud.