August 31, 2016 Joe Intrakamhang, Deanna Rowe Solutions Engineers Google Cloud Platform Overview Demo’s galore 

Confidential & Proprietary Google Cloud Platform 2 Joe Intrakamhang Solutions Engineer Chicago Email: [email protected] Deanna Rowe Solutions Engineer Chicago Email: [email protected] 

01 Introduction Google as a Cloud Service Provider 

For the past 16 years, Google has been building the world’s fastest, most powerful cloud infrastructure on the planet. 

Confidential & Proprietary Google Cloud Platform 5 2012 2014 2002 2004 2006 2008 2010 Google Research Publications referenced are available here: http://research.google.com/pubs/papers.html The Datacenter as a Computer: An Introduction to the Design of Warehouse-Scale Machines, 2009 http://research.google.com/pubs/pub35290.html GFS MapReduce Bigtable Google research in data technologies Colossus Dremel Flume Megastore Spanner Millwheel F1 PubSub 2016 Borg 

Confidential & Proprietary Google Cloud Platform 6 2012 2014 2002 2004 2006 2008 2010 GFS MapReduce Bigtable Google research in data technologies Colossus Dremel Flume Megastore Spanner Millwheel F1 PubSub 2016 Borg Tensor Flow 

Confidential & Proprietary Google Cloud Platform 7 2012 2014 2002 2004 2006 2008 2010 GFS MapReduce Bigtable Google research in data technologies Colossus Dremel Flume Megastore Spanner Millwheel F1 PubSub 2016 Borg Tensor Flow ? ? 

02 Google Technical Infrastructure 

Confidential & Proprietary Google Cloud Platform 9 “[Google's] ability to build, organize, and operate a huge network of servers and fiber-optic cables with an efficiency and speed that rocks physics on its heels. This is what makes Google Google: its physical network, its thousands of fiber miles, and those many thousands of servers that, in aggregate, add up to the mother of all clouds.” - Wired Images by Connie Zhou 

Google backbone We lay our own cables across the ocean Google confidential │ Do not distribute 

Global Network World’s Largest SDN Edge locations in 33 countries 100+ peering locations Global CDN Global Load Balancing with single Anycast IP Seamlessly autoscale to over 1M QPS with no pre-warming 

Sustainability Google datacenters have half the overhead of typical industry datacenters Largest private investor in renewables: $2 billion generating 3.2 GW Applying Machine Learning produced 40% reduction in cooling energy 

13 10 more coming by end of 2017 Google Cloud Regions Oregon Tokyo 

03 Cloud Platform Services 

© 2016 Google 15 Cloud Storage Cloud SQL Cloud Datastore Big Table Storage Compute Engine (IaaS) App Engine (PaaS) Container Engine (powered by Kubernetes) Compute BigQuery Pub/Sub Dataflow Big Data Dataproc Datalab Translate API Cloud Vision Speech API Machine Learning Cloud ML 

04 Compute 

Compute Engine App Engine Container Engine Continuum of Compute Virtualized hardware Abstracted computing power Manages your container cluster and actively schedules your containers Infrastructure at Google speed Build your scalable app faster 

High-performance Virtual Machines Consistently performant, scalable, highly secure & reliable. (Really) Pay for what you use We bill in minute-level increments so you don’t pay for unused computing time, and automatically apply sustained use discounts. Fast, Easy Provisioning Quickly deploy large clusters of virtual machines with intuitive tools. Compliance & Security All data written to disk in Compute Engine is encrypted on the fly and then transmitted and stored in encrypted form. Compute Engine Batch Run short duration, heavy compute jobs. The more flexibility in timing and location you give us the better the pricing! $ 

Confidential & Proprietary Google Cloud Platform 19 Compute Engine Demo 

Confidential & Proprietary Google Cloud Platform 20 Live migration for transparent maintenance (cont’d) 

Declarative management Declare your containers’ requirements, such as the amount of CPU/memory to reserve and keepalive policy, in a simple JSON config file. Container Engine will schedule your containers as declared. Better ops Your container cluster is equipped with capabilities, such as logging, container health checking, and autoscaling, to make application management easier. Docker support Container Engine supports the common Docker container format. And with Google Container Registry, Cloud Platform makes it easy to store and access your private Docker images. Managed container cluster Spin up a managed container cluster of virtual machines, ready for deployment. The vm nodes that comprise your cluster are fully managed, ensuring they are healthy and updated with critical patches. Container Engine Cloud flexibility With Red Hat, Microsoft, IBM, Mirantis OpenStack, and VMWare -- and the list keeps growing -- working to integrate Kubernetes into their platforms, you’ll be able to move workloads, or take advantage of multiple cloud providers, more easily. 

Powerful built-in services Managed services, such as Task Queues, Memcache and the Users API, let you build any application. Deploy at Google scale You can scale up to 7 billion requests per day and automatically scale down when traffic subsides. Focus on your code Let Google worry about database administration, server configuration, sharding & load balancing. Popular languages & frameworks Write applications in some of the most popular programming languages, use existing frameworks and integrate with other familiar technologies. App Engine Familiar development tools Use the tools you know, including Eclipse, IntellIJ, Maven, Git, Jenkins, PyCharm & more. Multiple storage options Choose the storage option you need: a traditional MySQL database using Cloud SQL, a schemaless NoSQL datastore, or object storage using Cloud Storage. 

23 Provision server and infra environments ● Install web and database servers ● Stitch it all together - db connection strings ● Duplicate for dev, test, and prod Manage the app ● Deploy it ● Scale it ● Pager duty What would it take to launch this tiny app? This is a 90s style guestbook. All it needs is a rotating GIF. 

24 Focus on code, not infrastructure 

25 Build efficiently with tools, services & APIs Integrated developer experience ● Integrated SDK across Cloud Platform ● Compile source into binary containers with one command Easy microservices ● Built-in API for building microservices App versioning and A/B testing ● Application versioning and traffic splitting for slow rollouts / rollbacks ● A/B testing across multiple experiments Common app building blocks ● Managed services and APIs for storage, queueing, caching, logging, and auth 

Confidential & Proprietary Google Cloud Platform 26 App Engine Demo 

05 Data Storage Solutions 

NoSQL SQL Blob Block Easy-to-use storage options 

Cloud Storage Cloud Datastore Cloud SQL Storage Store and manage data using a fully-managed, relational MySQL database Powerful, simple and cost effective object storage service Managed, NoSQL, schemaless database for storing non-relational data 

Durable Reduced Availability Storage Durable Reduced Availability (DRA) Storage has lower cost and lower availability than Standard Storage but is designed to have the same durability and performance as Standard Storage. Standard Storage Standard Storage provides the highest level of durability, availability and performance of all Google Cloud Storage services. It’s specifically designed for use cases requiring low latency and frequent data access, such as website content distribution and video streaming. Standard storage is all about performance. Simple Access Google created three simple product options to help you improve the performance of your applications while keeping your costs low. These three product options use the same API, providing you with a simple and consistent method of access. Cloud Storage Nearline Storage Nearline Storage is a low-cost, highly durable storage service for data archiving, online backup, and disaster recovery. Unlike other cloud or on-premises archive and backup offerings, you don’t have to wait hours or days to retrieve or access your data. 

ACID Transactions Cloud Datastore provides ACID transactions using optimistic concurrency control. Your application can execute multiple datastore operations in a single transaction in which either all succeed or all fail, ensuring the integrity of your data. Managed Database Cloud Datastore is fully managed. Google automatically handles sharding and replication in order to provide you with a highly available & consistent database. Autoscale with your users Cloud Datastore automatically scales depending on your needs. This allows you to focus on building your application and not on worrying about provisioning & load anticipation. Built-in Redundancy With a single API call, data is automatically replicated across multiple data centers. High availability & durability are built in from the very core. Cloud Datastore Schemaless Access with SQL-like querying No need to worry about data models and migration. Cloud Datastore is a schemaless storage service that allows you to be agile by removing the need to think about the underlying structure of the data. Access your data from anywhere Build solutions that span App Engine and Compute Engine, and rely on Cloud Datastore as the integration point. With the RESTful interface that is exposed by Cloud Datastore, data can easily be accessed by any deployment target. 

Control It's easy to manage and access your instances through a web Console or a command-line interface. Transfer data to your instance by importing and exporting databases and CSV files. Flexible Charging If you're running a lightly or sporadically used database, you'll save money by only paying for the time you access your data. Easier Migration; No Lock-in Standard connections and tools such as mysqldump, MySQL Wire Protocol, and JDBC make it easier to migrate onto (or off!) Google Cloud Platform, and avoids lock-in. Familiar Infrastructure Build and deploy for the cloud faster because Cloud SQL offers standard MySQL databases, the most popular open source database in the world. Instances available up to 16GB RAM, 100GB storage. Cloud SQL Fully managed No worrying about replication, patch management or database management: we take care of it. Security, Availability, Durability Your data is replicated in many geographic locations as standard, and failover between them is handled automatically. We also manage your backups, making it easy for you to restore when needed, including point-in-time recovery. 

What is BigQuery? Externalization of Google Dremel Convenience of SQL Petabyte-Scale and Fast Fully Managed, No-Ops Data Warehouse 

Confidential & Proprietary Google Cloud Platform 34 BQ Demo 

Fun BigQuery Stats Streaming ingest at peak Largest Data Lake on BigQuery Largest query by data size Largest query by rows 10.5 Trillion rows 2.3 Million rows per second 38 Petabytes 2.1 Petabytes 

Most importantly, complexity is hidden from end user We only paid $20 We just rented 9000 cores from Google for ~45 seconds 1 2 3 BigQuery - explained 4 End users do not need to think about cores 

06 Machine Learning 

TensorFlow ● Deep Learning technology currently powering over 100 Google services ● Generalizable to vision, sound, text, video and other data ● Runs on CPUs or GPUs, desktop, server, or mobile computing platforms. ● Distributed via Apache 2.0 OSS license 

What Cloud Machine Learning Can Do ● Fully managed service ● Train using a custom TensorFlow graph ● Batch and online predictions at scale ● Integrated Datalab experience ● Regression and classification tasks ● Runs on Google Brain using GPU & TPU 

Fully trained, easy to use Machine Learning models Cloud Translate Cloud Vision Cloud Speech Cloud Natural Language Stay tuned… 

Cloud Vision API Insight from images with our powerful Cloud Vision API 

Cloud Vision API Label Detection Landmark Detection OCR Logo Detection Face Detection Explicit Content Detection 

Confidential & Proprietary Google Cloud Platform 43 Vision API Demo 

Cloud Speech API Speech to text conversion 

Cloud Speech API Recognize Speech Streaming Recognition Transcribe Audio Transcribe stored audio Global Supports > 80 languages 

Cloud Natural Language API For sentiment analysis and entity recognition in a piece of text 

Confidential & Proprietary Google Cloud Platform 47 Identify Entities Entities with salience Cloud Natural Language API - Call API from anywhere, with support for embedded Text, and Google Cloud storage - Support for English, Spanish and Japanese Understand Sentiment Sentiment Analysis Text Syntax Analysis Identify POS with dependency trees 

Confidential & Proprietary Google Cloud Platform 48 Demo Speech & NLP 

Cloud Translate API Dynamically translate between thousands of available language pairs 

Confidential & Proprietary Google Cloud Platform 50 Translate API Features Language Detection Simple Integration Translate Many Languages 90 different languages, from Afrikaans to Zulu. Used in combination, this enables translation between thousands of language pairs. Translate API can automatically identify languages with high accuracy. Easy to use Google REST API. No need to extract text from your document, just send it HTML documents and get back translated text. High Quality Translations High quality translations that push the boundary of Machine Translation. Updated constantly to seamlessly improve translations and introduce new languages and language pairs. 

06 Closing