Contact us on 0121 314 0001 to discuss your SAN requirements.
This is We Solve IT's Recommended implementation of a Storage Area in a Virtualised Environment.
- Our solution of provides total redundancy if either a host or the SAN fail
- Due to the replica on the host’s local storage, you can recover from a SAN failure in under an hour
- With Virtual Server Hard Disks on the SAN in the event of Host failure VMs can be moved to other hosts while issues are resolved
- Maintenance on hosts can also be done with no downtime with Live Migration of Virtual Servers to other hosts
- Rolling replicas also allow for backup of files and settings, you can recover specific files and settings from the date of your choosing without full restore or any downtime
Storage Area Networks explained
This article offers an insight into the basics of traditional Storage Area Networks (SANs). However if you are looking for information regarding storage in a virtualised environment you will find our Storage Virtualisation page useful. Conversely, whilst an efficient Storage Area Network is a good start to a Disaster Recovery (DR) strategy, our DR services are discussed in more detail on our Disaster Recovery page.
Data is the underlying resource of which all computing processes are based, and such data is often a unique company asset. Since the turn of the century We Solve IT have been part of the continuous development of data storage over networks and as such have implemented cutting edge solutions for businesses. If you would like us to evaluate your requirements and to discuss which data storage network solution is a best fit for your company please contact us on 0121 3140001.
The storage requirements of companies today have grown exponentially over the last decade. To ensure company productivity the data needs to be easily and quickly accessible but also highly secure, reliable and configurable. The fundamental purpose of Storage Area Networks (SANs) is to transfer company data between their computer systems and storage elements.
The earliest approach to a storage network is Server Attached Storage (SAS). Whilst once highly popular it has now become quite unfashionable as more and more companies are moving away from this solution. SAS is facilitated by storage devices which are tightly coupled with the server bus using an adapter card. SAS usually relies on the Small Computer System Interface (SCSI) protocol, but as the storage is local to the server the preferred SCSI transport is usually Parallel SCSI. The significant limitation of this solution is its scalability. The number of storage devices that can be attached to the host server are limited by the parallel cables meaning that at certain points the introduction of additional storage devices require the implementation of new host servers. This has obvious cost implications as companies grow and data levels increase. Moreover, as additional servers are introduced the stored data is not easily shared between hosts. The functionality and manageability of the SAS approach is also limited by the difficulty of adding or removing storage without downtime implications.
As the cost and complexity of SANs reduced at the beginning of the century this led to an increase in their deployment by businesses. Storage Area Networks (SANs) have overcome the limitations of SAS. SANs provide a way to add storage without the downtime and disruption associated with server-attached storage upgrades. They also facilitate centralise data management, which greatly reduces overall operating costs compared to the aging SAS solution.
Business leaders are aware of the capital costs associated with implementing storage solutions as well as the operational expense in their maintenance. High-end disk drives, such as Fibre Channel (FC) drives, often account for a major portion of the total cost of storage networks. Naturally, businesses are now looking for cheaper alternatives to Fibre Chanel drives but obviously do not want to compromise the performance of their storage systems. The good news is that whilst the demand for storage solutions are increasing, the cost per Gigabyte is reducing. Consequently it is becoming more affordable for businesses to implement Storage Area Networks (SANs) and retain the company’s data for a longer period of time. The deployment of SANs are increasing dramatically, and this growth is expected to continue into the foreseeable future. There is a wider range of storage technologies which are available today as illustrated below.
- Serial Advanced Technology Attachment (SATA) is a strong contender when offering an alternative solution to Fibre Channel storage devices. The low cost and high performance of SATA plus the fact that it is a scalable solution has resulted in its popular use by enterprises. Many organisations have implemented SATA for storage applications such as backup and archiving as well as high bandwidth applications, such as video streaming and video editing. SATA satisfies the need for an inexpensive secondary enterprise storage solution whilst providing some high-end disk system characteristics. SATA is not the ideal solution for every organisation, for many enterprises especially those with mission-critical and production applications other alternatives are more viable. Whilst SATA disks are significantly cheaper than Fibre Channel Disks (less than half the cost) FC disks are technically far superior. SATA disks are best suited to reside at the “near-line” or secondary location within the networked storage hierarchy. SATA has slower drive mechanisms and limited interface functionality meaning that the levels of Input Output Per Second (IOPS) and MBps performance is inferior to FC drives.
- Serial Attached SCSI (SAS) drives superseded the older Parallel SCSI (Small Computer System Interface) bus technology that first appeared in the mid-1980s. It was designed to offer a cost effective alternative with improved performance. SAS is continuously evolving and improving and is now arguably more reliable than FC drives and less expensive. It uses a point-to-point architecture that transfers data to and from SCSI storage devices by using serial communication. Utilising dual-domain and dual-path configurations SAS can provide effective solutions for enterprises looking for a higher level of redundancy, reliability, and increased data availability within the storage network. SAS drives, similar to Fibre Channel drives, are designed for the rigors of enterprise use and heavy loads. They also safeguard data integrity via their comprehensive verification/error correction capabilities. A significant benefit of SAS is its flexibility. SATA drives can be plugged directly into SAS drive without any modifications or changes. The controller in the array and the software support a system with a mix of SATA and SAS.
- Solid State Drives (SSDs) offer outstanding performance but have less storage capacity per drive than hard disk drives (HDDs). Another limitation of SSD’s is the cost implications as HDDs are much more affordable. The main physical difference between traditional HDDs and SSDs is that HDDs are spinning disks whilst SSDs utilise silicon memory chips and have no moving parts. When determining whether the additional cost of utilising SSD’s is a price worth paying, enterprises should consider the extent to which high performing SSD’s would benefit the organisation. For businesses whose workloads do not require such high performance, investing in SSD drives may not be cost effective.
- Fibre Channel disk drives have been the standard data storage solution within organisations for more than a decade. The advantage of Fibre Channel drives are the levels of Input Output Per Second (IOPS) and MBps performance. They utilise cutting-edge technologies to maximize rotational velocities and data transfer rates, while lowering seek times and latency. In addition, the Fibre Channel interface provides robust functionality to process multiple I/O operations concurrently of varying sizes in both directions at once. Since 2009 the FC dive paradigm has begun to shift in favour of other solutions such as Serial Attached SCSI (SAS).
Connectivity within the SAN
Identifying which are the best storage devices to utilise within a SAN is only part of the problem. The data which is stored on these drives needs to be retrievable quickly and efficiently without any data loss over the network. Continuous connectivity is no longer a “nice to have” but is an essential requirement of modern businesses. What is also of utmost importance is the protection of the data that is stored, not only is important to retain data for the interest of the organisation, in many circumstances it is a legal requirement. Below is a selection of industry standards which are each worth consideration in overcoming the issues mentioned above.
- Redundant Array of Independent Disks (RAID) is a concept used to minimise the risk of loss of data over a network. It offers protection again the loss of data if a storage device fails. Historically this problem has been overcome by regular manual backups taken from the storage device. Whilst this solution was better than no attempt at all to back up data, any data inputted between backups could be lost if the storage device failed. RAID controllers maintain the storage devices in synchronisation by reading and writing (input/output (I/O)) to and from the disks. The result of this is that if one disk fails the data has already been replicated on another disk. It is therefore apparent of the significant importance of utilising RAID controllers in the SANs of all organisations.
- The set of standards known as Small Computer System Interface (SCSI) was set in the mid 1980’s but is now becoming less popular. SCSI allows personal computers to communicate with peripheral hardware such as storage devices, and is the conventional, server-centric method of connecting storage devices. Parallel SCSI allows multiple storage devices to be connected to host servers using a cable-based bus. The maximum length of the cables in a SCSI system is 25m and the maximum data transfer rate achievable on is 80 Mbps. A maximum of 15 devices can be connected to a SCSI bus at any time meaning that scalability becomes difficult and expensive. It is this restriction that has seen the demise of SCSI in favour of faster and more efficient solutions.
- Internet Small Computer System Interface (iSCSI) is a transport protocol that carries SCSI commands from an initiator to a target over an internet connection. It provides the mandatory performance and reliability for business-critical data and applications, and also has the potential to lower the costs of deploying networked storage. The benefit of iSCSI is that it enables the implementation of IP-based SANs, allowing businesses to use the same networking technologies, for both storage and data networks. iSCSI is suited to run over almost any physical network which reduces the need for additional network technology which would only be used for storage. It has had significant success in the SME sector but has had limited impact on large-scale organisations. Its success has been achieved by leveraging the ubiquity and cost efficiency of Ethernet.
- Traditional Fibre channels are networks of copper or optical cable offering a low latency, reliable high-speed solution over long distances, of up to 50 miles. It was developed to overcome the limitations of the SCSI infrastructure. It works well in storage networks but as it was designed to support large, complex networks with millions of devices the infrastructure is now considered too complex and too expensive. Many believe that the use of Fibre Channels (FC) will soon diminish as Ethernet is now far superior and the fibre channel is no longer able to compete.
- InfiniBand is an industry standard, channel-based, switched-fabric, interconnect architecture for servers. It has the characteristics of a bus technology and has always stayed well ahead of other network technologies. It provides reduced server-to-server latency from milliseconds to microseconds and has high bandwidth and extremely low CPU overhead with very high price/performance ratios. Many believe that Infiniband will become prevalent in the future however it is yet to be embraced by the market on a large scale. InfiniBand switches meet the demands of businesses with its economical storage fabric. Interconnects such as InfiniBand can scale a storage system efficiently by utilising network based connectivity from the front end. Infiniband is economical and has higher bandwidth capabilities than Ethernet, however to install Infiniband would likely require companies to upgrade their network infrastructure.