How can organisations ensure their branch and remote office storage doesn’t leave them out on a limb?  

BRANCH AND REMOTE OFFICES are the core of most businesses. But this is particularly true in the retail, manufacturing, financial services and hospitality sectors. In these industries, the majority of customers meet and engage with companies at branch offices rather than the corporate HQ - and certainly not at their data centres! Because of this, branch and remote offices need fast and reliable access to critical customer-facing applications and highly sensitive customer data. When these applications run well, customers receive the service that they expect, but if they fail (due to a fluctuating broadband or WAN connection to the HQ or data centre) customers notice straight away.

It might be something as seemingly simple as the inability to take a card payment, retrieve loyalty card details or accept a gift voucher, but it means that a company’s reputation takes a massive hit, and the likelihood of repeat custom drops substantially. mAs such, critical customer centric applications and associated data have to be available at all times - even when companies are under pressure to simplify and reduce the cost of their IT generally and data storage particularly. One way they are trying to achieve this objective is to replicate the cost effective virtualised storage environments that have taken hold in at HQs and data centres in branch and remote office locations.
Protecting the downtown office from downtime
The pressure is on to reduce the storage costs associated with supporting remote applications while ensuring there is little or no downtime of business critical applications at branch offices or remote sites . To achieve this goal, organisations need to deliver high availability to eradicate downtime on business critical applications. Downtime can be particularly damaging to businesses because it directly equates to lost sales. As an illustration, a study by CA Technologies found downtime costs all European organisations with more than 50 employees a total of more than Ä17 billion in revenue each year.
In addition to e-commerce, downtime inhibits cash register operation, credit card processing, inventory management and many more aspects of physical branch operations. All of which take a toll on total sales. The effects of downtime can be particularly profound in multi-site infrastructures. Typical branch office IT infrastructures consist of six to eight business applications requiring a small amount of servers and storage. But the applications running in these environments are often business-critical, so the data has to be always available.

Keeping IT local
One way organisations could seek to alleviate the threat of downtime would be to centralise those applications by moving them to the cloud. But in this model, any outage would run the risk of affecting multiple locations. Keeping applications local would ensure any downtime only affects a single location and minimises the overall risk. In addition, certain applications need to remain at the edge as they have requirements that cannot be met through the cloud or at distance, such as performance, latency and bandwidth or the data has to reside locally for compliance reasons. If organisations opt for the local approach, they need to try and deploy the minimal amount of hardware on each site to minimise the capital costs associated with hardware acquisition and reduce ongoing operational costs related to power and cooling. Branch locations are also subject to floor space constraints and to IT equipment in non ideal environments, so they need to maintain a small IT and data storage footprint.
However, IT management still needs to be centralised because most branch offices do not wish to have the cost of a dedicated, local IT resource. For many organisations, the ideal would be an implementation they could set and forget, a simple, automated deployment that would provide them with a plug-and-play solution.
Server virtualisation allows companies to achieve this, but while virtualisation drives the consolidation of servers, it also increases the need for high availability which, in turn, requires shared storage. Traditionally, shared storage has been delivered via a storage area network (SAN), but it is not ideally suited for branch locations for a number of reasons:
 External storage arrays are single points
of failure;
 SAN adds complexity to the infrastructure;
 It requires specialist staff to manage
storage locally;
 SAN ties the user to a vendor and
depreciates in value, requiring eventual
replacement or upgrade;
 SAN has a high CAPEX and OPEX cost.
How to achieve high availability in branch offices
There are a number of system characteristics that need to be addressed to ensure high availability for branch and remote offices:
 The infrastructure must have no single
points of failure. This requires a minimum
of two servers and shared storage per
location with provision for avoiding “split
brain”
 Application performance is critical. Poor
performance associated with accessing
data over a slow network link to a
centralised or cloud-based application
could lead to unacceptable processing
delays, contributing to a poor customer
experience. Locating the data and
processing closer to users and
customers, would remove delays while
improving application performance and
customer experience.
 The system must maximise server and
storage utilisation. This is important for
any organisation trying to do more with
less, but especially so in branch and
remote office environments where
a minimal amount of IT and storage
infrastructure needs to support multiple
disparate workloads.
 The infrastructure must keep its IT
footprint to a minimum. This is difficult
to accomplish, especially in organisations
with tens or hundreds of branches. Only
a small amount of floor space is devoted
to back-office and IT infrastructure, and
the environments are harsh for IT
equipment because there are usually few
controls such as clean power, dust
filtration and air conditioning.
 Small changes to infrastructure design
can have a significant effect on overall
capital and operational costs, as they
are multiplied by the number of locations.
External SAN or NAS arrays are excessive
for the relatively low storage capacities in
edge environments and bring
complexities with them.
 The IT solution should be quick and
easy to deploy in all edge sites.
The simpler the solution and the fewer
infrastructure components, the faster it is
to deploy into production and the less risk
of downtime before deployment.
 The solution should be hardware-
independent, ensuring an organisation
can employ multi-vendor purchasing
policies, avoid vendor lock-in and take
advantage of competitive pricing.
 Ease of management is critical. Multi-
site organisations do not have skilled IT
employees at branch offices and remote
locations, preferring to centralise the IT
function at the head office or data centre.
This makes monitoring and management
more difficult as the number of sites
increase. A centralised management
capability can add tremendous value by
enabling skilled employees in central
areas to administer the solution remotely.
Why virtualised shared storage has the edge
A more advanced method for organisations to eliminate downtime and ensure high availability of business-critical applications at the edge is to use virtualised shared storage. It eliminates single points of failure by delivering highly available synchronously mirrored storage across multiple servers using internal direct attached disk devices. Should one server fail, the remaining server continues to support the applications until a new server is provisioned.
By providing simple design and flexible deployment options, it is possible with solutions such as StorMagic’s SvSAN to enable two-server configuration at each edge site, reducing operational and capital expenses while optimising availability and uptime. Virtualised shared storage can also optimise performance and utilisation by using internal server and storage and solid-state disks to accelerate I/O requests.

In addition, active-active clustering ensures all nodes are working to serve I/O. Planned and unplanned downtime can be minimised because hardware replacement and upgrades can take place during office hours while the system is online. Deployment is simple and repeatable with software deployments or upgrades scripted from a central management location and initial configurations handled from a central location. This means virtualised shared storage can be a plug and play solution at the edge site. Automating the deployment process also reduces time to value significantly and eliminates human error.
Branching out with IT
Many organisations face significant everyday business challenges, but IT and storage shouldn’t be among them, especially in remote and branch offices. To ensure operations run smoothly and systems remain online as much as possible, organisations need to deploy solutions that simplify the infrastructure, minimise spend and maintenance, enable applications to perform and bring maximum value.

Virtualised shared storage is a robust, highly available solution that addresses all the requirements of the edge and helps organisations ensure maximum uptime and operational stability.