Plugging the gap in Cloud resiliency

Cloud Imperative
Growing numbers of SMEs are opting to put core business applications in the cloud. The costs are compelling; the flexibility is appealing: when it is possible to start up a new business with nothing more than a printer, an IP phone and an Internet connection delivering both access to cloud based applications and VoIP calls – why invest in any on premise equipment?

There is just one, not so small, flaw in this model. What happens when the Internet connection fails? Too many organisations are over confident about the resilience and reliability of standard ADSL connections. Most will insist that they have relied on email for years and never had a major problem. But how true is this assertion? A short term email outage is far from unusual – but the business can cope for an hour or two.

But how does the shop remain open when there is no access to the cloud based EPOS system or stock information? Or the accountancy practice that has created a cloud based client service model? As one retailer recently reported, every single month around 25 of its 425 retail outlets endure ADSL outages – failure to put resilience in place would result in untenable business loss.

How will the business access core information? Receive critical in-bound calls from customers and suppliers? Will staff have to be sent home? Just how long could the business survive without access to information, email or, in some cases, telephone?

Single Point of Failure
So what are the options for achieving resilience? The obvious choice for organisations that rely on a leased line or ADSL would be to run a second, backup line. But what path is the resilient ADSL line taking from the office? Is it using the same copper connection infrastructure; and heading for the same exchange? The likelihood is that even if the business opts for a different provider, the underpinning infrastructure will still be provided by BT – if there is a problem at the exchange or the cable is damaged, both lines, primary and secondary, will fail simultaneously. Not so resilient after all.

The alternative is to have a dedicated line to a secondary exchange. The problem is that this is not an option for many businesses – there is no other exchange nearby. Even where it is an option, the cost will run into the £10,000s. Again, hardly feasible for most companies.

So what are the alternatives? Mobile? While many companies have eschewed the use of mobile due to the perceived lack of bandwidth, growing numbers of organisations are now exploiting business grade services that can use the 3G network to deliver resilience. While users should see little difference in performance between the ADSL and a standard 3G connection, for organisations that require higher bandwidth it is possible to use a bonded 3G/4G connection that combines between two and to six 3G/4G SIM cards to deliver upwards of 20Mbps download speeds.

Multiple Layers of Resilience
Not only can the solution be set up to failover in seconds in response to ADSL problems; but the business can opt for additional resilience by using multiple network operators. Should any problem arise with the primary operator, the system will failover to the next one – and so on. With multiple levels of resiliency, the business can feel completely confident in its ability to retain access to core information and services at all times.

Indeed, while ADSL is prone to competing with domestic, or non-business traffic, resulting in slow broadband speeds during times of high consumer activity – most notably evenings and school holidays; business grade networks do not carry domestic traffic. Even if the 3G network does get busy with business traffic, the system can intelligently switch to another, less busy carrier.

For the end user, the seamless failover from the primary connection to 3G within 15 or 20 seconds will be transparent. In addition, whichever network operator is being used, the company retains the same IP address – there is no need to reprogram the mail server or VPN. The solution can be tailored to reflect business specific needs – whether using Ethernet or ADSL as the primary connection, the different connections at remote locations or the need to include remote users on tablets into the secure environment.

Furthermore, companies can also opt to limit the way the resilient line is used. For example, those opting for a single, rather than bonded, mobile data connection may want to limit voice calls to inbound to ensure customers can always get in touch. A bonded 4G connection can – and does – support an entire office irrespective of call volume or the number of cloud services routinely accessed.

In addition, a company doesn’t have to use the mobile connection only for resilience – some businesses are opting to exploit the two connection model to route mission critical and less important data and activity down different connections. With this approach, the resilient failover is always in place, but the business is also gaining continuous value from the DR investment.

Conclusion
Moving to the cloud makes sense on many levels – but if it is not supported with robust and effective resiliency and contingency plans the ramifications can be extremely severe. Expecting 100% uptime is simply not reasonable. With the business now dependent upon rapid access to cloud based data and applications it is time to take your head out of the sand and consider the best way to deliver resilience and contingency. Repairing a broken ADSL line typically takes between five and ten days – a lack of resiliency could cost the business a lot more than a few days inconvenience.