Revenue Management & Assurance – Can revenue management be a revenue maker? Oct 2015 issue of VanillaPlus
Will CSPs remain locked-in to lock-in?
The expectation is that the move to virtualisation will put an end to vendor lock-in in telecoms. It’s true that software defined networking (SDN) and network functions virtualisation (NFV) are set to eliminate the need for function-specific telecoms network hardware which historically has seen vendors locked-in or obliged to remain with the same set of vendors.
Understandably, traditional network equipment vendors aren’t making haste towards virtual networks. The traditional business of shifting costly boxes in their thousands to near-captive customers is too lucrative and too attractive to accelerate the death of. Yet such vendors recognise that the requirement they place on CSPs to make huge investments every decade in function-specific hardware is unsustainable. To take a farming analogy, their thinking is that if they continue to sell function-specific equipment to CSPs, they’ll have a highly profitable result, similar to butchering a sheep, but they will only be able to do that once because the sheep will be dead.
Instead, they’re re-aligning their businesses to enable them to shear the CSP sheep on an ongoing basis, taking management fees and software licensing subscriptions alongside sales of basic hardware. As far as virtualisation goes in terms of network equipment, its management stack and associated services, this approach of shearing rather than killing, is viable.
However, the virtual world is software-controlled and software-defined. It is also based on an IT foundation and that brings us to the IT hardware vendors – the server guys. They’re licking their lips at the potential market that CSPs represent.
They’ve never farmed sheep before, but there’s a growing appetite for the slaughter.
A server in the virtual network, in spite of what has been promised, is not just a dumb box with storage and computing power. It needs to be controlled and managed to adopt the functions necessary to manage network traffic and those instructions need to be issued in a language the dumb box understands. Whether by accident or design this universal language – a network esperanto – is yet to emerge.
I accept that it’s early days but protocols such as OpenFlow, Netconf and OpenDaylight remain immature and IT vendors’ proprietary approaches open up a very real danger that CSPs will be faced with trading their locked-in relationships with network equipment makers for ones with big IT equipment vendors
For lock-in to be ended, a universal key is required but the companies that could provide it are only offering locks that require their own, expensive and bespoke laser cut keys. I fear the lock-in is far from over, it is just changing key-holder.
Enjoy the magazine!