This week in Google Cloud — “GKE 1.7 is here, London calling, a Transfer Appliance, and much improved TCP congestion control”
Google Cloud Platform news over the past couple of weeks included :
- a new Datacenter in London to run your applications and store your data, all with Google’s reinforced commitment to General Data Protection Regulation (GDPR) compliance across GCP. Three more regions in the EU are also coming soon!
- less than a couple of weeks after the release of Kubernetes 1.7, Container Engine now runs this newer version and focuses on better stateful app support, more security features, and networking support for hybrid deployments. Other features include alpha GPU support, auto-repair and auto-upgrade of nodes, new cluster auto-scaling strategies, and an updated admin UI. Check it out!
- The Transfer Appliance, a rackable high-capacity storage with capacity of up to one-petabyte compressed is now available in beta.
- In the spirit of giving GCP customer access to the same technology Google uses for its users we give you TCP BBR congestion control, a technology documented in a technical article last December.
Continuing with the “The network is the computer” department :
- How Google wants to rewire the Internet (nextplatform.com)
- The Regional Performance Problem (youtube.com)
- Tools for Profiling Network Performance (youtube.com)
This week’s Big Data news covers Apache Airflow, Dataflow, BigQuery, but also Genomics and our latest Data Engineering Coursera courses :
- Straggler elimination with dynamic re-balancing & auto-scaling in Cloud Dataflow (infoq.com video)
- [Podcast] What’s Next for Apache Beam? Featuring Frances Perry of Google (talend.com podcast)
- How to aggregate data for BigQuery using Apache Airflow (cloud.google.com)
- Using R with Google BigQuery, Part 2 (storing and retrieving data frames) (cloud.google.com)
- DNA Sequence Alignment as a Cloud Service (medium.com)
- Data Engineering on Google Cloud Platform: Coursera courses (medium.com)
From the “Containers, Kubernetes, …” department :
- Google Container Builder Part 1 (Cloud Rolling Update) (youtube.com)
- Istio and Kubernetes — Kelsey Hightower (youtube.com)
- Kubernetes: Understanding Pods vs. Containers (by Tim Hockin)(speakerdeck.com)
- Open Container Initiative (OCI) Releases v1.0 of Container Standards (opencontainers.org)
- Free Intro to Kubernetes by the Linux Foundation (edx.org)
From the “we’re stronger and smarter with partners” department :
- Going Hybrid with Kubernetes on Google Cloud Platform and Nutanix (cloudplatform.googleblog.com)
- How we’re collaborating with Citrix to deliver cloud-based desktop apps (blog.google)
- Announcing Managed Google Cloud Platform by Rackspace (blog.rackspace.com)
From the “customers and users talk best about GCP” department :
- Using GCP for massive drug discovery virtual screening (Silicon Therapeutics) (googleblog.com). Another strong case for preemptible GCE instances.
- Google not Amazon. Make fantastic savings in a server-less world (3wks.com.au). A case for looping in the CFO.
- From AWS to GCP, TABLEAPP Architecture Story (slideshare.com). GKE and cost analysis.
- BQ: Powering mobile innovation (with GKE, BigQuery, Cloud SQL & more) (cloud.google.com)
- Real time data processing using Google Cloud and OpenTSDB (blog.greta.io). Too many pieces to list, check out the architecture diagram.
- From Kafka to BigQuery: A Guide for Streaming Billions of Daily Events (medium.com). Compares data loading to BigQuery via batch and via the streaming API.
- Moving Thumbtack’s data infrastructure to Google Cloud Platform (cloud.google.com). From queries taking 1h+ to complete to Cloud Storage + Cloud Dataproc (Spark) + BigQuery.
- Marvin, NYTimes’ go-kit server for Google App Engine, is now open source (github.com)
From the “what would a weekly recap be without Machine Learning updates?” department :
- Facets: An Open Source Visualization Tool for Machine Learning Training Data (opensource.googleblog.com)
- Announcing the winners of our Machine Learning Startup Competition (blog.google)
- Building Your Own Neural Machine Translation System in TensorFlow (research.googleblog.com)
From the “ICYMI” department :
- How to get started with Cloud Spanner in 5 minutes (cloudplatform.googleblog.com)
- Help keep your Google Cloud service account keys safe (cloudplatform.googleblog.com)
- Using Clusters for Large-scale Technical Computing in the Cloud (cloud.google.com)
- NOAA Global Surface Summary of the Day Weather Data (cloud.google.com)
- Scaling Dedicated Game Servers with Kubernetes: Part 2 — Managing CPU and Memory (compoundtheory.com)
- BigQuery Mate chrome extension to add some extra functionality to the user interface (by Mikhail Berlyant) (chrome.google.com)
- Universal Now: Now, on Every Cloud (zeit.co)
- Catatumbo 2.2.0 (the persistence framework for Cloud Datastore) is now final (catatumbo.io)
If you haven’t caught up with the GCP Podcast recently, you probably missed three very fine episodes :
- #86 Broad Institute and Platinum Customers with Lukas Karlsson and Mike Altarace
- #85 Istio with Varun Talwar and Sven Mawson
- #84 All about Kaggle
The pic of the week is said to be an example of the dynamic nature of protein target and is taken from the Silicon Therapeutics story above :
That’s it for this week!
-Alexis

