Most people see a product as an inanimate object, but to a Product Manager like me, they are much more than that. Products give a soul to organisations’ – they breathe life into every department from procurement, finance, project management, service delivery and service support. So when I got the opportunity to spend some time with the network guru’s working on the ANS service desk a few months ago, I jumped at the chance.
Seeing how a product is working within a customer’s environment, how it lives and breathes, enables or disrupts a business was more valuable than I could have hoped. In the 6 weeks, I spent working alongside some of the most talented, tenacious network engineers I think I have ever come across, I learned one major thing.
If the network is down it’s a big deal. Simple. It can directly affect a business’s financial performance.
It was during this 6 weeks that I gained first hand exposure to a common, real-world networking problem. A customer was experiencing an issue with a flapping link. The circuit would be stable for a few moments, fail, then restore again a few seconds later.
We have a whole team of Cisco qualified engineers who work tirelessly to resolve issues in the quickest time possible; because if the network isn’t performing properly, then applications aren’t running. And if applications aren’t running then there can be huge impact to the business, so what did they do?
The engineers shut down the primary link to force traffic across to the secondary link. Service was restored – albeit with manual intervention. They then worked with the carrier to isolate the problem and restore stable service to the primary circuit, a faulty fiber being the culprit.
I sat and thought about this. In a traditional WAN design, this type of fault can result in loss of service, the primary circuit may not be down long enough for the back-up to take-over. And what if the back-up circuit has an unknown fault? I know of plenty of Service Managers who have got hot under the collar when they find out that the backup circuit they have convinced the CTO to buy for every site, doesn’t work because nobody has thought to insist on regular failover testing!
A traditional WAN can’t make intelligent decisions based on parameters such as packet loss, without regular failover testing you may not know the back-up is live – but there is a solution that can. In a traditional WAN design, this type of fault can result in a loss of service as the primary circuit may not be down long enough for the back-up to take over. A traditional WAN can’t make intelligent decisions based on parameters, but thankfully, there is a solution that can – SD WAN.
SD-WAN is an overlay network that sits on top of a traditional network (the underlay). By using mechanisms to sense the quality of the path and techniques for load sharing and traffic re-assembly, it can automatically “route” around these types of circuit instability issues in real time. In the case of the customer with the unstable connection, the overlay would simply route around this using the existing standby circuit and then re-use the primary circuit once stable again. The customer and their users would never know – and the service desk team could work busily in the background restoring their primary service.