GCP 2015/2016 Updates @ GCPUG Hanoi

Transcript

1. Ian Lewis Developer Advocate, Google Cloud Platform GCP Updates GCPUG Hanoi
2. Confidential & Proprietary Google Cloud Platform 2 Ian Lewis Developer Advocate - Google Cloud Platform Tokyo, Japan google.com/+IanLewis-hoge @IanMLewis
3. Google Cloud Platform Images by Connie Zhou
4. Confidential & Proprietary Google Cloud Platform 4 Cloud Bigtable Q2 2015 Q3 2015 Q4 2015 Q1 2016 GKE GA Dataflow GA Tensorflow Cloud Vision API Innovation
5. Confidential & Proprietary Google Cloud Platform 5 Networking Management Networking Big Data Storage Machine Learning Developer Tools Compute Engine Container Engine Compute Google Cloud Platform Compute
6. Confidential & Proprietary Google Cloud Platform 6 Compute Engine • Custom Instance Types • Change size of volume without downtime • Persistent disks up to 64TB • Preemptable Instances GA
7. @tekgrrl #gcpug #googlecloud Greek for “Helmsman”; also the root of the word “Governor” • Orchestrator for Docker containers • Supports multi-cloud environments • Inspired and informed by Google’s experiences and internal systems • Open source, written in Go Manage applications, not machines Kubernetes
8. @tekgrrl #gcpug #googlecloud • Ready for Production since 1.0 • Speed and Stability improvements • Deployments • Ingress • ConfigMaps • Horizontal Pod Autoscaler • New Dashboard UI Kubernetes 1.2
9. @tekgrrl #gcpug #googlecloud • GA (Aug 2015) • Managed Kubernetes • Manages Kubernetes Uptime • Manages Updates • Cloud Logging • Use Google Cloud VPN to have cluster IPs coexist with your private network IPs Google Container Engine
10. 10 • Part of HTTP load balancer backends • Cache data at the edge of Google’ s network Google Cloud CDN
11. Confidential & Proprietary Google Cloud Platform 11 Management Networking Compute Big Data Storage Machine Learning Developer Tools Google Cloud Platform Big Data Cloud Dataproc Cloud Dataflow BigQuery
12. Confidential & Proprietary Google Cloud Platform 12 BigQuery • Long Term Storage Pricing • Query Explain • User Defined Functions (UDFs)
13. Confidential & Proprietary Google Cloud Platform 13 BigQuery Query Explain
14. 14 • GA (Aug 2015) • Stream + Batch Processing • Apache Beam Google Cloud Dataflow
15. 15 • GA (Feb 2016) • Managed Hadoop, Spark, Pig, and Hive • Create data managed data processing clusters in a few clicks Google Cloud Dataproc
16. Confidential & Proprietary Google Cloud Platform 16 Management Networking Compute Big Data Storage Machine Learning Developer Tools Google Cloud Platform Storage Cloud Bigtable
17. 17 • Beta (May 2015) • Supports HBase API (Managed HBase) • The same database that powers Search, Analytics, Maps, and Gmail. • Read/Write p99% latency <6ms Google Cloud Bigtable
18. Confidential & Proprietary Google Cloud Platform 18 Google Cloud Bigtable
19. Cloud Translate API Cloud Prediction API Cloud Vision API Fully trained, easy to use Machine Learning models
20. Confidential & Proprietary Google Cloud Platform 20 Vision API • Classify Images by Tag • Detect Faces and Sentiment • Detect inappropriate content • Text Detection (OCR) • Landmark & Logo Detection
21. Confidential & Proprietary Google Cloud Platform 21 Vision API
22. Confidential & Proprietary Google Cloud Platform 22 Vision API
23. Cloud Speech API
24. TensorFlow www.tensorflow.org
25. Confidential & Proprietary Google Cloud Platform 25 Tensorflow • General Purpose Machine Learning Library • Python (Core in C++) • Models can run on many devices (even Android!) • Use models in production using Tensorflow Serve • Distributed version that uses gRPC
26. # define the network import tensorflow as tf x = tf.placeholder(tf.float32, [None, 784]) W = tf.Variable(tf.zeros([784, 10])) b = tf.Variable(tf.zeros([10])) y = tf.nn.softmax(tf.matmul(x, W) + b) # define a training step y_ = tf.placeholder(tf.float32, [None, 10]) xent = -tf.reduce_sum(y_*tf.log(y)) step = tf.train.GradientDescentOptimizer(0.01).minimize (xent) Example
27. TensorBoard: visualization tool
28. github.com/woodrush/neural-art-tf Neural Art in TensorFlow http://www.anishathalye.com/2015/12/19/an-ai-that-can-mimic-any-artist/
29. Cloud Machine Learning Announcing
30. Fully managed, distributed training and prediction for custom TensorFlow graph Supports Regression and Classification initially Integrated with Cloud Dataflow and Cloud Datalab Limited Preview - cloud.google.com/ml Cloud Machine Learning (Cloud ML)
31. ALPHA Use your own data to train models GA GA Cloud Storage Google BigQuery Cloud Machine Learning Cloud Datalab BETA Develop / Model / Test
32. Thank You