First, users need to decide what kind of storage fits their needs. There are various parameters to consider:
- Persistency: Users need to consider whether they need a persistent or non-persistent storage option. This is the concept of how secure the user needs the data to be stored.
- Type of data: Different data types require different storage; types include database files, videos, images, photos, and text.
- Performance: Another factor is what kind of performance will be needed, what capacity they need, and how many input/outputs per second will be required.
- Durability: This pertains to how many copies a user will want to maintain (how many times should the data be replicated).
- Connectivity: Users might desire local or network storage.
- Protocol: This will determine the kinds of applications a user can build.
There are various types of storage services provided by the Oracle Cloud Infrastructure:
- Local NVMe: This is a locally attached storage with super high performance, giving large amount of IOPs. Local NVMe is available in the availability domain.
- Object storage: This type of service is advantageous for having truly persistent, durable storage that extends beyond the lifetime of an instance itself. In this, data is managed as fixed sized blocks. So, firstly user needs to make a partition to create a file system and then mount them. This is also available in availability domain. object storage is an internet scale, high performance storage platform. here data is managed as object. It's good for unstructured data like logs, videos, pictures, text files. Object are stored in a bucket which have their own unique name within tenancy. It is multiple storage tiers and have a private access from OCI resources (i.e., compute). All the object stores go in a flat hierarchy within service. A top-level container which contains all the buckets objects called namespace and it have a unique name. Objects can also be automatically versioned let say from version v1 it can automatically upgraded to version v2. Data is encrypted by default and have their own encryption key. There is a feature called Auto-tiering in which it can move from one tier to another tier automatically.
It has 3 Tiers:
1. Standard Storage Tier (Hot): It is fast, immediate and frequent access and can be retrieve instantaneous, strongly consistent. It can also update any time gives the most recent copy of data.
2. Infrequent Access Storage Tier (Cool): It's ideal for data which is not much frequent. Data has to be stored for a minimum of 31 days. Retrieved fees is there when user get the data back.
3. Archive Storage Tier (Cold): It is for the data which is not need right now and have a minimum retention period is 90 days. Data in this tier need to be restore which takes minimum 1 hour and need to be downloaded which take min. 24 hours. Archive bucket cannot be upgraded. - File storage: It is a shared file storage system which can read and write but storage is shared. As there are two compute instances which need to have a shared storage system availability domain. Here, user can manage data as files and directories. File storage is hierarchical collection of documents organized into name directories. There are many use cases for File storage: E-business Suite, micro service and container, scale out app and analytics etc. As in OCI, user have data center in OCI region which have a couple of compute instances and they all are talking to shared file system. It supports NFS v.3 distributed file system.
- Block Volume: It is storage for web. This type of storage can be used to store photos, videos, log files, text file etc. User can simply access this data using internet client who accesses these objects using simple HTTP verbs like PUT and GET. We have 4 block volume tiers:
It has 3 Tiers:
1. Basic Tier: Good for large sequential I/O Workloads, such as streaming, data warehousing. Has 2 IOPS/GB
2. Balanced Tier: Balanced choice for Random I/O, things like your boot disks and it have up to 60 IOPS/GB.
3. Higher Performance Tier: which is for the most I/O-demanding Workloads and have up to 75 IOPS/GB.
4. Ultra Higher Performance Tier, which is for the highest I/O-demanding Workloads and have a 90-225 IOPS/GB.
There are some features of block volume:
- 1. Block volumes are persistent disks that can be attached to compute instances.
2. Block volumes provide persistent and durable storage. user can create, attack, detach, delete the disks and it can even keep data after deleting the instances.
3. They are in-transit encrypted can use bring your own keys.
4. Many VMs can read and write data from a single block volume.
5. A user can also resize block volumes on demand.
6. Replication is also possible in block volumes, and they are grouped together for easy management.
- 1. Block volumes are persistent disks that can be attached to compute instances.
- Archive Storage: It is used to store the data that is accessed infrequently and require longer retention period. It is more cost effective than object storage for preserving cold data. It also used bucket as logical container for storing objects.