Ubiquitous Data Access – Your Storage In The Cloud

The increasing needs of organizations to have ubiquitous access to always available reliable storage has presented significant challenges to IT organizations to ensure availability, reliability and security when using legacy physical data center resources. To accommodate for these needs, storage virtualization offers on-demand rapidly expanding storage in line with changing business needs and ensures redundancy for highly available solutions in private, public and hybrid deployments.

Storage Virtualization

Storage virtualization provides for the creation of a logical resource pool of grouped physical heterogeneous storage devices, offering on-demand scalable resources as a single logical pool that increases utilization of underlying physical storage, improves centralized management of storage provisioning, increases redundancy, availability and allows for extension and migration without interruption (Li, et al., 2005).

(allcadsolutions, 2020)

Storage virtualization involves removing reliance on and abstracting the functions of the physical storage units and devices, concealing it from the host application, servers and consumers.

Storage Virtualization Schemes and Architectures

Methods for storing

Object storage is an implementation that manipulates and manages data structures and storage as distinct objects within a container which is not typically grouped in hierarchical folder and file structures. To represents the object or data structure, combined with any metadata that describes the file, generated or manually captured by the user. It is given a unique identifier by which the object can be retrieved or manipulated. Manipulation normally takes the form of full replacement and is typically offered via web-based communications.

File based virtualization makes use of NAS devices to group and pool resources to make it easier to manage and utilize. It masks the complexity of combining and managing physically separate appliances in a pool opaque to and consumable by users, applications and hosts. As a shared network resource pool, it requires the management of proper access rights, locking and security measures. File storage is typically managed in hierarchical folder structures to allow for grouping, securing and collocating of files. File storage also includes limited metadata such as creation and last update date, file sizes and ownership, but can be extended by the operating system. Its file and folder structure also negatively impact performance as the numbers increase.

Block storage is managed in fixed size chunks of data, where files can be spread across many such blocks, depending on the size. This is as opposed to Object and File storage that manages the item is a whole. Within the blocks, the data is ordered, but for multi-block items, the blocks do not have to be sequential, and the underlying storage system is responsible for reconstructing the item on request. Block storage stores limited metadata and must be enriched at an application layer. In many cases, block storage is the base or foundation layer for File and Object storage and is typically deployed in a SAN environment.

(Brown, 2017)

Storage Technologies

DAS (Direct Attached Storage) systems are devices that are directly connected to the host and offer low cost and low maintenance block storage (Surianarayanan & Chelliah, 2019). Because it is directly linked to the host, it can offer far superior performance over NAS, but lack the ability to be shared across hosts.

NAS (Network Attached Storage) offers shared management across many physical storage devices served to many hosts. Its central approach offers better oversight and management, reduced costs and improved utilization and redundancy (Surianarayanan & Chelliah, 2019). NAS in its simplest form is a file system provided over a network and can offer improved redundancy, reliability and availability.

SAN (Storage Area Network) offers high performance block level storage which transfers block level data from the storage across the network to the hosts (Surianarayanan & Chelliah, 2019). It offers the combined benefits of the speed of block level DAS and the availability, sharing and flexibility of NAS systems. SAN systems require large cost investments as these systems require upfront implementation, redundant storage and network design.

(Nguyen, 2020)

Storage Virtualization Implementations

Host-Based storage virtualization requires additional software on each host to manage logical volumes, which abstracts the complexities of the underly storage topology from the host (KAY, 2008). This layer is responsible for intercepting the calls and IO requests and mapping it to the underlying devices. In many modern operating systems these Logical Volume Managers are built in.

Network-Based storage virtualization requires a dedicated device or server to manage heterogenous underly storage devices, can offer caching of data for improved performance and can facilitate replication across underlying devices (KAY, 2008). It is independent of the host or underlying system and requires no additional compute form the hosts to manage the logical mappings. At this point it is important to consider in-band or out-of-band in relation to the management of metadata and data paths for fabric virtualization.

Storage-Based virtualization offers advanced disk arrays with features such as replication, cloning and snapshots but does not necessarily provide support across heterogenous storage devices as each vendor might have their proprietary implementation. Typically a dedicated appliance manages the pooling and metadata of resources such as RAID controlers.

(Peglar, 2011)

Cloud Storage

Cloud storage is a service model that offers remote reliable and highly available storage across the internet for on-demand use, which is highly scalable and cost effective (Nagarajan & Suguna, 2014). Cloud storage is acquired from a third-party vendor that manages the underlying resources pooling, capacity, security and durability of the storage. It provides just in time cost management and scaling, reduced upfront and total cost of ownership, increased deployment times and improved business continuity.

Offering Comparisons

Many cloud service offerings exists, ranging from standalone services focused on storage only to integrated cloud services which is managed from within a cloud management portal alongside many other cloud services, and include features such as File Management, Collaboration capabilities, and Secure Administration through Roles and Responsibilities based Identity and Access Management. They range in costing models from limited size free subscriptions to tiered models. These vendors include offerings such as Google Drive, Dropbox, OneDrive, Apple iCloud, Citrix FileShare, Amazon Drive and the integrated solutions such as Amazon’s S3, EBS and EFS, Azure’s Blobs and Azure Files, and Google Cloud Storage, Persistent Disk and Cloud Filestore.

A comparison for the various cloud managed storage services is listed below.

Object Storage (Soni, 2020)

File Storage (Soni, 2020)

Block/Disk Storage (Soni, 2020)

References

allcadsolutions, 2020. Virtualization Solutions. [Online] Available at: https://www.allcadsolutions.com/virtualization/ [Accessed 04 Apr 2020].

Brown, L., 2017. Object Storage, Should I Be Using It And Step By Step To Deploy It?. [Online] Available at: https://www.whatwouldlukedo.com/whats-object-storage/ [Accessed 04 Apr 2020].

KAY, R., 2008. Storage Virtualization: This primer will help you understand the technology's origins, market, benefits and challenges. COMPUTERWORLD, 42(40), pp. 34-37.

Li, B., Shu, J. & Zheng, W., 2005. Design and Implementation of a Storage Virtualization System Based on SCSI Target Simulator in SAN. Tsinghua Science & Technology, 10(1), pp. 122-127.

Nagarajan, B. & Suguna, J., 2014. A Review on Cloud Data Storage in Virtual Perspective. International Journal of Computer Science and Information Technologies, 5(5), pp. 6027-6031.

Nguyen, A., 2020. NAS and SAN Introduction, Introduction to RAID, and Introduction to Solid State Drives. [Online] Available at: https://sites.google.com/site/andreanguyencis4397/module-1/nas-and-san-introduction [Accessed 04 Apr 2020].

Peglar, R., 2011. Storage Virtualization I What, Why, Where and How?. [Online] Available at: https://docplayer.net/8731679-Storage-virtualization-i-what-why-where-and-how-rob-peglar-xiotech-corporation.html [Accessed 05 Apr 2020].

Soni, V., 2020. Top public cloud storage providers in 2020 (with comparison). [Online] Available at: https://www.dailyhostnews.com/top-public-cloud-storage-providers [Accessed 05 Apr 2020].

Surianarayanan, C. & Chelliah, P., 2019. Essentials of Cloud Computing. s.l.:Springer.