GCP Updates | Semtember 30, 2019

Posted on Oct 9, 2019


Compute Engine – name change for memory-optimized VMs: GA

To more clearly identify VM instances as members of the memory-optimized machine-type family, all n1-ultramem and n1-megamem VMs have had their “n1” prefix changed to “m1”. Documentation

Compute Engine – persistent disk scheduled snapshots: GA

Use snapshot schedules to regularly and automatically back up your Compute Engine workloads as well as your zonal and regional persistent disks. And define how long to keep snapshots with a snapshot retention policy. Documentation

いままでなかったのか 😳


AI Platform – custom containers: GA

Run your apps in a Docker image by building your own custom container to run jobs on AI Platform. Use machine learning frameworks and versions – as well as non-ML dependencies, libraries, and binaries not otherwise supported on AI Platform. Documentation

中ではCloud Runが使われてんのかな 🤔

AI Platformの方のドキュメント見るとDockerのインストールとかDockerfileの書き方とかまで書いてあっておもしろい。

AI Platform – use predefined machine-type names to configure training job: GA

Gain greater flexibility when allocating computing resources for training machine learning jobs by using names of certain predefined Compute Engine machine types. You can also customize how your job uses GPUs when training with TensorFlow or using custom containers. Documentation

Cloud Video Intelligence API – logo detection: beta

Detect, track, and recognize the presence of over 100,000 brands and logos in video content. Documentation


Cloud Data Loss Prevention – new Google Cloud Platform Console user interface: GA

Gain greater visibility and control over sensitive data with the new Cloud DLP UI. Inspect and classify data in Cloud Storage, BigQuery, and Cloud Datastore repositories with a few clicks. Manage inspection jobs, create templates, and examine findings in your console. Documentation

Cloud Asset Inventory – Organization Policy and Access Policy support: GA

View and assess more resource and policy types in one centralized, managed inventory service with the newly onboarded policies. Documentation

Identity Platform multi-tenancy support: beta

Create unique silos of users and configurations within a single Identity Platform project to set up data-isolation boundaries. The silos can represent different customers, business units, subsidiaries, or other identifying characteristics. Documentation


GCP Marketplace Private Pricing: beta

Create custom quotes for prospective customers for your Kubernetes apps or integrated SaaS solutions with GCP Marketplace Private Pricing. The quote gives customers the ability to purchase your product on the GCP Marketplace at the negotiated price in their GCP Console. Documentation