Storage System Technologies and their Configurations

In order to efficiently handle your data and acquire the best storage option for your enterprise data, it is necessary to develop an understanding of the available solutions and methods of data storage. This article is meant to facilitate this purpose and enable you to develop a clear understanding of the three storage levels and the three configurations for each of these storage levels. These storage levels and configurations are capable of addressing variable enterprise data storage requirements.

In other words, these storage solutions facilitate the creation of a comprehensive, reliable and efficient storage solution.

The Three Storage Levels: File, Block & Object

Storage levels are basically three ways of how data is handled when storing it. Each storage level stores data differently and the stored data is utilized and accessed differently. Based on these differences, each storage level best suits a specific kind of data storage requirement.

As depicted in the image, there are three types of storage:

• File Level.
• Block Level.
• Object Level.

File Level Storage

File storage is a hierarchical way of storing files. An individual file is accessed using the path or address of the file. Attributes associated with the file such as its owner, who has access to the file and size of the file are stored in the metadata of the file system. It’s a simple system to use and implement. In comparison to block level storage, file level storage is inexpensive to maintain and use. File Storage is seen and deployed in Network Attached Storage (NAS) systems.

The NAS storage stores the files and hosts them over a local network constituting a file sharing system with reduced latency as compared to cloud storage solutions. If the data requirement of the enterprise is to share files between departments located on the same premises or building, then a NAS is a better option than enterprise cloud storage. This is because NAS hosts the files over a local network, keeping the data close and easily accessible as compared to cloud storage.

Block Level Storage

Block level storage divides the data into chunks called blocks. The collection of multiple chunks leads to the generation of a specific file. A block is accessed by generating a SCSI (Small Computer System Interface) call to its associated address. This introduces micromanagement of the storage space. The other difference between file level and block level storage is that in file level, data also has the associated metadata.

In block level, the only relevant metadata is the address to the block and arguably even that is not exactly metadata. Individually, a block has no description or ownership related to it; hence it makes little sense by itself. It only makes sense when combined with other chunks of data like itself to form a complete file or data.

This granular level storage is deployed in Storage Area Network (SAN) environments. Besides the difference of how file level and block level handle data, there’s also a fundamental difference in the utilization of the two. NAS appliances, that utilize file level storage, are used only for file storage and sharing. It cannot be used for any other purpose. Meanwhile, SAN storage facilitates individual formatting of file systems like NFS, NTFS or SMB (Windows) or VMFS (VMware). It can be used to either store files or as a database, virtual machine file system and more.

Object Level Storage

Objects combine the data with its metadata. This bundle or object is assigned an ID. The ID is generated versus the content of the object. This ID is used to call or access the object when needed. File level storage stores files in a hierarchical structure; however, object storage utilizes flat structure to do so. An object within this architecture can be called using this ID regardless of the location, it can either be in the local storage or stored in a geographically separate location. The beneficial feature unique to object level storage is that it facilitates users to define the attached metadata of an object. This enables vast opportunities for data analytics and efficient storage management. Object level storage is used in cloud backup such as Microsoft Azure and Amazon Web Services (AWS S3, S3-IA, Glacier etc.).

Configurations of each Storage Level/Mode

The above mentioned storage levels can be configured in three ways depending on the enterprise data requirements.

• Standalone configuration.
• HA (High Availability) configuration.
• Scale out configuration.

Standalone Configuration

As the name suggests, there’s only a single chassis involved. The chassis has hard drives, SSDs, RAID controllers making it a complete storage.

StoneFly Integrated SAN Storage products are an example of this configuration mode. It is an easy to use ISC 12GS-Class Integrated Storage Concentrator that facilitates a cost effective iSCSI and provides an optional Fibre Channel solution. This solution can be used for small, medium and even large enterprises in primary or secondary, main data center, remote sites and/or branch offices. This product family offers large storage capacities at low prices and is available in 12GB SAS or SSD combinations of 4, 6, 8,12,16,24 or 36 drives per base unit. It also offers the most cost effective expansion via 12-bay, 16-bay or 24-bay expansion units for up to 256 drives. And, the enterprise backup appliance also facilitates cloud connect services. This implies that while storing data on the appliance, the enterprise can also opt to backup to Azure or AWS or to StoneFly Private Cloud Storage.

 

High-Availability (HA) Configuration

 

This configuration ensures high availability by having twin systems working simultaneously. This prevents downtime or disruption in cause of a single system failure as the other node continues to work.

StoneFly provides two appliance families in this configuration: Voyager and Flash Voyager DX Series. These appliances come with a fully modular design with no single point of failure. The appliances are equipped with two clustered redundant storage controllers. Each storage controller starts with 16 scaling up to 44 Parallel storage core engines. This modular design is extremely robust and scalable to meet any demanding workload. And as with the previous appliances, these appliances also facilitate cloud connect services. This means, backup jobs can keep copies on the local infrastructure and enable enterprises to backup to AWS (Amazon Web Services) or Microsoft Azure.

Scale out Configuration

In this configuration, it’s not one or two storage controllers; it’s many of them. This implies that if you need improved performance or more storage capacity, you acquire multiple nodes.

The above image depicts how scale out works. Initially, there is a single node and with configuration you can have multiple nodes and keep on adding them; this is scaling out. All of the nodes within the system, aggregate together. This architecture, as additional nodes are being added, has a multiplicative affect for performance. For instance, if the enterprise was experiencing speed issues with a single node; with the addition of another node, the speed is doubled. Similarly, if another node is added then the speed is tripled and so on and so forth.

referred to as scaling up. With StoneFly’s appliances, you get to scale out. Scaling out is the addition of additional nodes. The advantage of node addition as compared to expansion units, this scaling out improves performance and capacity due to the workload being distributed among the nodes.

Summary

There are three types of storage levels: File, block and object. File level storage is simplistic and hierarchical, in which data is called using the address. NAS appliances employ file level storage. Block level creates chunks of data, these chunks individually don’t make sense unless combined with other chunks to constitute the complete data. The only metadata involved in Block level storage is the address of the chunks of data and even that is arguably not considered as metadata. Object level storage creates objects by combining data and metadata. An ID is assigned to an object and is used to call to it when needed. SAN (Storage Area Network) employs block level storage whereas cloud backups utilize object level. There are three configurations that these storage types or levels are available in: Standalone, HA and Scale out. Standalone comprises of a single node that can be expanded via expansion units. High Availability (HA) comprises of twin storage controllers that ensure continuity of operations even when one of them fails. Scale out configuration facilitates redundancy and increases performance and capacity also improving workflow and productivity. An enterprise can devise their perfect storage solution using a specific storage architecture in a specific configuration. In order to do that, enterprises need to identify their data requirements. Only then can an enterprise make an efficient and comprehensive backup and disaster recovery plan.