From Google Cloud: Ice Cold Archive Storage

From Google Cloud blog, by Dominic Preuss, and Dave Nettleton, directors, product management

Storage provides the foundation for enterprise infrastructures, and that’s particularly true with cloud, where readily available data storage that’s cost-efficient is a must.

At Google Cloud, we think you should have a range of straightforward options to store your data and reliably access it when and how you need it, with the performance you need.

However, moving your data to the cloud isn’t a one-size-fits-all process.

We develop storage tools for flexibility, so you can make moving to cloud as seamless as possible. Here’s what we’re announcing:

• Introducing a new archive class for Cloud Storage (coming soon)

• Making Cloud Filestore available with higher throughput

• Regional Persistent Disks generally available next week

• Bucket Policy Only and V4 signature support for Cloud Storage now in beta

• Cloud IAM roles now available for Cloud Storage Transfer Service

Storage that’s optimized for your business
Having the right storage for all your types of data can add efficiency and performance, plus save your business money. We announce some new options for data storage needs to grow with you over time.

Introducing new class of storage for data that’s ice cold
We will be rolling out an entirely new archive class of Cloud Storage designed for long-term data retention. Available later this year at price points starting from $0.0012/GB/month ($1.23/TB/month), the archive class is intended for data that would probably otherwise be stored in tape archives. Unlike tape and other glacially slow equivalents, we have taken an approach that eliminates the need for a separate retrieval process and provides immediate, low-latency access to your content. Access and management are performed via the same set of APIs used by our other storage classes, with integration into object lifecycle management so that you can tier cold objects down to optimize  TCO. Data in Cloud Storage is always redundantly stored across availability zones with 11 9’s annual durability. As with other storage classes, data stored in the archive class in multi-regional locations is stored geo-redundantly to maximize durability and availability. Contact sales to plan your archive migrations.

Making cloud Filestore generally available with higher throughput
Many traditional applications require shared file storage. We announce the availability of Cloud Filestore, our managed file storage system that’s built for high performance. Cloud Filestore’s premium instances will now provide increased read performance up to 1.2 GB/s throughput and 60,000 IO/s. Customers have been using Cloud Filestore when they need to collaborate on shared files, or need a high-performance cache, or have to move an application that uses file semantics to the cloud. Create a file share now and learn more in the documentation.

Regional Persistent Disks are available next week
Regional Persistent Disk provides active-active disk replication across two zones in the same region. Regional Persistent Disks deliver durable storage that is synchronously replicated across zones and are a great option for high-performance databases and enterprise applications that also require high availability. You don’t have to worry about manually replicating your data across different zones in a region. Learn more in the documentation.

A primary benefit of moving to the cloud is the performance improvement. You can add control and protection over your cloud workloads and get the performance improvement benefits. We’ve increased the throughput limits per instance for Google Persistent Disk Standard to 240MB/s for writes, and 240MB/s for reads. If you’re running throughput-driven workloads on GCP, you can now benefit from 33% higher per-instance throughput limits for Persistent Disk Standard. Learn more in the documentation.

Number of Persistent Disks that can be attached to a VM has increased to 128
Now, all machine types with at least one vCPU will be allowed to attach up to 128 Persistent Disks. Shared-core and burst machine types will be limited to 16. Learn more in the documentation.

Adding more safeguards for cloud data storage
We protect cloud data throughout its lifecycle, and we’re always working to add more safeguards to cloud storage.

Bucket policy only for cloud storage
Having the right set of controls for the cloud also means adding security controls at the right level to help carry policies and governance into your cloud infrastructure. You can now manage permissions centrally and uniformly. Bucket Policy Only for Cloud Storage (now in beta) lets enforce Cloud Identity and Access Management (IAM) policies at the bucket level for consistent and uniform access control for your Cloud Storage buckets. Administrators can also ensure that all newly-created buckets utilize Bucket Policy Only through a new organization policy setting. Enforcing uniform Cloud IAM policies at the bucket level can help prevent accidental public exposures of your organization’s objects. Learn more in the documentation.

V4 signature support for cloud storage
V4 signatures (now in beta) provides improved security and enables customers to access multiple object stores using the same application code. In addition to the security gains, this also simplifies multi-cloud and hybrid cloud deployment and adoption for our customers. In addition to HMAC keys, V4 signed requests are also supported for Google RSA keys. Learn more in the documentation.

Cloud IAM roles available for Cloud Storage Transfer Service
In addition, Custom Cloud IAM roles and permissions in Cloud Storage Transfer Service are now available. This allows security and IT administrators to use Cloud IAM permissions for creating, reading, updating, and deleting transfer jobs. These permissions can be assigned directly to individual users or to specific roles.

Whatever your workload, from backing up an image archive to crunching a genome, we’ve got you covered with cloud storage tools and services that make it easy to migrate and run on GCP.