Wednesday, 9 October 2013

SIP Based Call Centres - 12 years on - Time for a light weight cloud approach to contact centre architectures

Back in 2001 - Wow that seems a long time ago now that I sit here in Oct 2013, I wrote a white paper which was published back then by the International Engineering Consortium in their Annual Review of Communications, called SIP Based Call Centres - A vendor independent architecture for multimedia contact centres and thanks to the Internet can still be found out there.

In the paper I sang the virtues of open standards and SIP for building contact centres capable of meeting the needs of a modern multimedia environment (voice, video and Instant messaging and presence). Since then (IMHO) SIP has become quite frankly bloated by extension after extension, and one SIP doesn't necessary work with another SIP, so SBC now have to mediate between different flavours.

I think its time to revisit signalling and communications in modern contact centre, after all the point is to let customers reach someone who can help them in whatever way best suites the customers need, in an efficient no frills way.

On reflecting on what I wrote 12 years ago and the current surge of interests in WebRTC and all things HTML and cloud, it struck me. Now is the time to rethink the way we build multimedia contact centres both for the customer and the business that uses them. Actually it wasn't too new a thought - I'd already put some top level architecture ideas together about a year ago when looking at a new approach to telecare with some colleagues over at Inmezzo. The approach for the next generation of contact centres means a much lighter footprint - no complex clients to use for the agents, no nasty plugins in the browser for customer, integration with the PSTN, voice, video and text chat from desktop or smart phone.

What doesn't change is the basic requirements for session queueing and routing, or the ability to present information about the contact.

So what does this new architecture look like.... Well I'll leave my thoughts on that for the next blog entry.

Monday, 7 October 2013

We are entering a new era of software defined communications

What now feels like a life time ago back in the dim and distant past ,OK - not so dim and distant 1980s, I was a young apprentice engineer with my training and interests firmly focused on Hardware (micro electronics more specifically). In 1987 I made the bold step towards software engineering as a choice, since it seemed, back then, that software was the way forwards to creativity and new solutions for communications. Having "converted" to software engineering by way of a degree in Computer Science at Aberystwyth University, I then continued in that vain becoming a professional software engineer. I still didn't forget the electronics that got me started, and this combined with software engineering put this to good use as a telecommunications engineer. Moving through the heady days of Nortel Passport Frame Relay Switches, Cisco AGS and MGS routers and Cabletron hubs.

Then the Internet happened.......

For the telecoms engineers out there (like me), in my humble opinion (IMHO) SDN is the networking equivalent of softswitches? The OpenFlow API is what MEGACO/SIGTRAN is to soft-switching and the so-called Orchestration layer the Stored Program Control Logic from the switch fabric and routing fabric. The media gateway is the Hardware forwarding platform (L2 switch fabric being physical of logical vSwitch, Microsoft Network Virtualisation and Virtual Subnet Identifiers), not to mention MPLS enabled VPLS and the rest of the MPLS "family".

And now in the telecoms field we're getting all excited about Network Function Virtualisation (NFV - . This is really about extending the Virtualisation platforms/techniques like those provided by VMWare ESX and Microsoft HyperV into the telecoms space and placing the Functional elements of the NGN network on these virtualisation platforms. As the ETSI White paper states:

"leveraging standard IT virtualization technology to consolidate many network equipment types onto industry standard, high-volume servers, switches and storage."

This allows the economies of scale that the enterprise networks are gaining through "cloud compute" to the telco space, by implementing these traditionally custom hardware based elements (such as media gateways, firewalls, SBCs, Routers, CSCFs) in standard (commodity Servers) architecture servers and a hypervisor layer, utilising the same virtualisation tools which enable flexibility of deployment of these elements in more standard environment, enabling features such as live migration of network functions under failure, multiple instances of a network function on the same hardware elements, and more interestingly (from my perspective) the ability to create multiple instances belonging to different customers (or even carriers) on the same hardware platforms.

Time to re-write the architecture rule book.