Getting to know Google Cloud Storage.
Building and operationalizing storage systems.
Google Cloud Storage stores unstructured data, such as data files, pictures, movies, archive data, and other types of data.
Google Cloud Storage is an online file-based storage service that allows you to store and access files on Google Cloud Platform. The service leverages Google's cloud's speed and effectiveness with sophisticated confidentiality and sharing features. It is an Infrastructure-as-a-Service offering on GCP.
It saves files in atomic format, meaning that when you view a file in Cloud Storage, you view the whole file. Therefore, it cannot be considered a file on a block storage device that enables users to browse and access distinct blocks within the file.
Managing Items in a Bucket
A bucket is a collection of objects with shared access restrictions at the bucket level. Users may, for example, have access to read from all objects in one bucket and write to all objects in another. Additionally, each object within a bucket may have different access restrictions.
All bucket names must be distinct since Google Cloud Storage employs a global namespace for bucket names. No need for uniqueness in object names. It is impossible to change the name of a bucket once created. Transfer the bucket's contents to a new bucket with the appropriate name; if you need to alter a bucket's name, then delete the old bucket.
Google suggests the following when choosing a bucket name:
Each bucket name must be globally unique; cloud storage throws an error when a proposed bucket name is currently in use.
Bucket names are publicly viewable. Thus, don’t use Personally Identifiable Information when naming a bucket, as anyone can inquire into the existence of a bucket name.
After deleting a bucket, anybody can use its name to create a new bucket.
A bucket name may be used in a DNS record as part of a CNAME or A redirect. (You must confirm your domain ownership to create a domain name bucket).
Lowercase letters, numeric characters, dashes (-), underscores (_), and dots are the only characters permitted in bucket names (.). There are no spaces allowed.
Storage classes
A storage class is a sort of metadata that is shared by all objects. The storage class assigned to an object influences its availability and pricing policy. For example, unless set differently, objects in a bucket inherit the storage class of the bucket when created.
The following are the four classes of Cloud Storage:
Standard
Nearline
Coldline
Archive
Each of these four classes has three different location types that affect their availability and pricing, with availability and durability increasing from regional to dual-region and finally multi-region.
All storage classes come with a 99.999999999% annual durability, unlimited storage with no minimum object size and low latency.
Muti-region: also called geo-redundant storage, reduces the chance of a regional disruption by retaining object clones in several regions. They help lower access time and latency by spreading data duplicates to places closer to the consumers.
Dual-region: users indicate two regions on the same continent of their choice. They help bolster performance and availability while accessing Google Cloud products in one related area.
Region: keeps additional copies of an object in different zones within the same area.
Standard Storage
Standard storage is ideal for regularly accessed data that are only held for a short time. There is no limit on the minimum storage duration, and no associated data retrieval cost.
Standard storage costs between $0.020 - $0.035 per GB-month for storage, depending on the object location and has a monthly availability of 99.99% in regions and > 99.99% in multi-regions and dual-regions.
Standard storage is suitable for providing website content, streaming media, serving interactive workloads, or supplying data for mobile and gaming apps.
Nearline Storage
Nearline storage is a reduced-cost, durable storage service for seldom accessed data. It is ideal for data accessed less than once in 30 days and where somewhat reduced availability and data access fees are appropriate trade-offs for reduced at-rest storage costs. There is a 30 days limit on the minimum storage duration and a charge of $0.01/GB for data retrieval.
Nearline storage costs between $0.010 - $0.020 per GB-month for storage, depending on the object location and has a monthly availability of 99.9% in regions and 99.95% in multi-regions and dual-regions.
Nearline storage is ideal for storing lengthy media, data archiving, and backup accessed less than once in 30 days.
Coldline Storage
Coldline storage is a more affordable, durable storage service for occasionally accessed data. It is best for data accessed less than once in 90 days and where somewhat reduced availability and high data access fees are appropriate trade-offs for more reduced at-rest storage costs. There is a 90 days limit on the minimum storage duration and a $0.02/GB charge for data retrieval.
Coldline storage costs between $0.004 - $0.007 per GB-month for storage, depending on the object location and has a monthly availability of 99.9% in regions and 99.95% in multi-regions and dual-regions.
Coldline storage is appropriate for backup or archiving data accessed less than once in 90 days.
Archive Storage
Archive storage is the most affordable, durable storage service for rarely accessed data. It is best for data accessed less than once in 365 days, where reduced availability and higher data access fees are appropriate trade-offs for well-reduced at-rest storage costs. There is a 365 days limit on the minimum storage duration and a $0.05/GB charge for data retrieval.
Archive storage costs between $0.0012 - $0.0030 per GB-month for storage, depending on the object location and has a monthly availability of 99.9% in regions and 99.95% in multi-regions and dual-regions.
Archive storage is appropriate for disaster recovery, compliance and legal reasons, accessed less than once in 365 days.
Data Retention and Lifecycle Management
Data has a life cycle that includes creation, constant usage, and occasional usage during which it is stored online, archived, and removed. Although not all data is transmitted across all stages, it is critical to address lifecycle problems when designing storage systems.
Data lifecycles and retention rules are typically not significantly affected by the architecture of the storage solution, but they affect how the policies are carried out. For instance, objects may be transferred from standard to nearline storage using Cloud Storage lifecycle rules after a certain amount of time.
Cloud Storage offers object lifecycle management guidelines that enable remote modifications to how objects are stored in a bucket. Such policies, which are allocated to buckets, define guidelines for modifying objects, and rules bind objects in such categories. The rules carry out lifecycle operations, such as removing an object and changing the storage class. Data lifecycle management may execute rules depending on the object's age, creation date, number of current models, and storage class.
Retention rules are yet another type of data governance regulation. A retention rule enforces object retention by employing the Bucket Lock function of the Cloud Storage bucket. Configuring a retention policy ensures that any object in the bucket or subsequent objects are not removed before they meet the retention rule's designated period. This function is very beneficial for meeting government or business compliance. In addition, a retention policy that has been enforced cannot be overturned.