Storage

Question 1

What is RAID?

Answer 1

Redundant array of independent disks that can be configured in a way to improve data redundancy and performance

Question 2

What are common RAID levels used?

“Zero Stripes, One Mirror, Five Parity, Six Double, Ten Combined”

Answer 2

RAID10 (1+0) - mirrored sets in stripe

RAID0 - striping (improved performance, no redundancy)

RAID1 (mirroring, full redundancy)

RAID5 Striping with distributed parity (good balance of performance and redundancy)

RAID6 Striping with double distributed parity (better fault tolerance than RAID 5)

Question 3

What is striping in RAID configurations?

Answer 3

Process of dividing data into smaller segments and spreading them across multiple drives

Question 4

What is an array in the context of RAID?

Answer 4

Collection of multiple storage devices to function as a single storage device. This approach is used to improve performance, increase storage capacity, or provide data redundancy

Question 5

In what situations is SDS a better approach than RAID?

Answer 5

1. RAID limited to single system or storage arrays while SDS can span multiple systems, data centers, or geographic regions
2. RAID is tied to specific hardware controllers while SDS is hardware-agnostic and can work with various storage types
3. RAID has fixed configurations while SDS has customizable policies for data distribution and protection
4. Raid is often managed individually while SDS has centralized management of multiple storage resources
5. RAID is limited to physical size of the array for scaling while SDS can easily scale but adding new storage resources

Question 6

What is software-defined storage (SDS)?

Answer 6

Approach to data storage in which the programming that controls storage-related tasks is decoupled from the physical storage hardware with RAID like functionality but with more flexibility

Question 7

Key features for SDS?

Answer 7

Abstraction: Separates storage management from physical hardware.

Automation: Enables automated storage management and provisioning.

Virtualization: Creates virtual storage pools from physical resources.

Question 8

In what situations is RAID a better approach than SDS?

Answer 8

1. RAID limited to single system or storage arrays while SDS can span multiple systems, data centers, or geographic regions
2. RAID is tied to specific hardware controllers while SDS is hardware-agnostic and can work with various storage types

3.

Question 9

What is block storage and some of its core characteristics?

Answer 9

“Block storage is a low-level storage abstraction that provides raw storage volumes to applications and operating systems.

Each block has a unique address allowing for direct and random access

Accessed through low-level protocols like iSCSI, Fibre Channel, or NVMe over Fabric

Low latency and high IOPs

Consistent performance, especially for databases and transactional workloads

Question 10

What is object storage?

Answer 10

Manages data as objects, each with a unique identifier.

Highly scalable and designed for durability and availability.

Objects include the data, metadata, and a globally unique identifier.

Ideal for unstructured data, especially in large volumes.
Commonly used for backups, archives, static website hosting, and as a data lake for analytics.

Typically accessed via RESTful APIs over HTTP.

Question 11

What is file storage?

Answer 11

Presents a hierarchical structure of files and folders.
Supports file-level operations and POSIX compliance.

Ideal for sharing data across multiple instances or users.

Well-suited for content management systems, development environments, and general-purpose file sharing.

Offers strong consistency and file locking mechanisms.

Question 12

What is NVMe and its benefits over using standard block storage?

Answer 12

nonvolatile memory express - higher throughput and faster response times

Question 13

What allows NVMe to produce higher throughput and faster response times?

Answer 13

Has it’s own set of protocols:

Question 14

SSD vd HDD

Question 15

Flash Storage?

Question 16

Instance vs block store?

Answer 16

Instance storage is physically attached on the disks on the underlying instance itself and are ephemeral. block storage like EBS are network attached storage where the data saved on the block store volumes are persistent

Instance store is much faster since you’re writing directly to disk on the instance

Question 17

Key differentiators for using block vs file vs object storage?

Answer 17

Key Differentiators:

Scalability: Object storage excels in scalability, followed by file storage, then block storage.

Performance: Block storage typically offers the highest performance for transactional workloads.

Metadata: Object storage provides rich metadata capabilities, file storage offers basic metadata, while block storage has minimal metadata.

Access Patterns: Block for random access, file for shared access, and object for sequential access and distributed systems.

Consistency: File and block storage typically offer strong consistency, while object storage often provides eventual consistency with options for strong consistency.

Cost: Object storage is generally the most cost-effective for large data volumes, while block storage can be more expensive per GB but offers better performance.