There is a lot of talk these days in the communications industry about how the days of hardware are finished and software is the only way to go. This seems to be a recurring topic of conversation no matter what the actual communications subject is. In this article we will look at the wider topic and how it relates to Unified Communications.
How has this potential shift to software come about?
In the past most telecommunications processing was done in dedicated hardware that was built using Digital Signal Processing (DSP) chips. These DSPs were dedicated devices that were optimised for specific functions such as video encoding, audio mixing etc. This approach ensured that there was sufficient processing power available to perform these complex functions, as generic computing power (such as that provided by your PC or a server) was insufficient for the task. Over time the capability of this off-the-shelf computing power has improved in line with Moore’s Law (computing power doubles every 18 months: en.wikipedia.org/wiki/Moore%27s_law) so that now software based solutions are viable for many functions.
You can see how this evolution has happened just by looking at the progress of software video clients running on a PC or laptop. First you had dedicated video boards (built on DSPs) that had to be installed inside a PC and worked in conjunction with a software program (anyone remember the PictureTel LiveLan?). Next step on from that was the Polycom ViaVideo solution – this used a dedicated hardware camera connected to the PC for video encoding and the software on the PC did the decoding (video encoding is about 10 X more processor intensive than decoding). Next evolution was the Polycom PVX software, which could run without dedicated hardware but delivered a limited resolution due to the available computing power. Finally we arrive at Polycom RealPresence Desktop that can deliver HD quality video using nothing but software (as long as your processor is powerful enough).
This is just one example of how we have seen the capability of software solutions advance, and deliver capabilities that could previously only be done in hardware. We’ve seen similar advances in all areas of technology. It’s happening at the heart of communications networks as core Telco class 4 & 5 switches are steadily being replaced with call processing software running on generic computing hardware. It’s happening even with the most basic of devices, as organisations look to potentially using soft phones and UC client software rather than a traditional hardware telephone for a user’s desk.
What does this mean from a Unified Communications perspective?
The flexibility of software solutions now opens up more options for both users and administrators of Unified Communications solutions. From a user perspective they now have the option of a fully functional VoIP or UC client deployed on their desktop, or even their smartphone. This ensures they have full communications capability irrespective of location or device, thereby providing significant opportunities for productivity improvement. It is now possible to be a connected member of your organisation irrespective of where you are - no matter whether you are on the road, travelling to a foreign country or simply at home, you have the ability to have voice, video and instant messaging all from software clients running on your device of choice.
From an administration perspective, software solutions in the core of a UC solution now enables infrastructure to be deployed on top of generic computing power that is probably already running in an organisation’s data centre. This potentially provides a lower cost of entry for UC solutions, as well as enabling organisations to have a centralised, consolidated computing platform to power all their communications solutions. Software also enables organisations to scale up their core UC solutions much quicker, as they can use spare generic computing power and resources rather than having to order dedicated DSP-based hardware with the associated costs and lead times.
Software sounds like it offers many advantages over hardware. Does this mean that hardware solutions are “dead” and software is the way forward?
Software solutions certainly have advantages, some of which are described above. In addition some collaboration technologies such as Scalable Video Coding (SVC) are well suited to a software infrastructure. As video processing and mixing is done at the endpoint with SVC rather than in the core infrastructure, it is well suited to a less powerful software core rather than a dedicated DSP hardware one.
In addition a software core has much more deployment flexibility than hardware. For example if an organisation has purchased MCU ports they can potentially deploy those ports on software bridges “spun up” in data centres that are geographically appropriate to the users. Taking this to its logical extreme, you could for example deploy ports in an Asia Pacific data centre, and then as EMEA comes online for business hours, port licences could be deactivated in APAC and redeployed to other regions as required. A video collaboration management tool that integrates with the virtualisation platform management tools could even automate this process.
However before we get too carried away with the potential of software we should remember that hardware solutions still have a place and will do for many years to come.
For example with VoIP and UC clients, many companies initially looked to have a phone free desk deployment, but have since reconsidered and now deployed physical handsets. The vast majority of users are still conditioned to the physical handset experience when communicating by voice and find a pure software experience quite alien. Many analysts have predicted that the desktop phone faces extinction but the reality seems to be different. People still want the handset when they are sat at their desk. This is summarised extremely well by Robert Arnold on the Visionary IT blog
“Despite over-hyped industry perceptions about the imminent and swift death of the desktop phone, this tried and true technology remains a viable market today and one that will remain in play for years to come.”
When we look at the core infrastructure there is still definitely a place for hardware as well as software. For organisations running multi-codec immersive telepresence solutions, or wanting to deploy full HD 1080p AVC solutions, then software simply cannot provide the necessary scale of infrastructure required. In many Asia Pacific markets there is still a requirement to create large-scale “town hall” type conferences with 100 or more endpoints all connected at 1080p into a single conference. The ONLY way to be able to achieve this in an effective and manageable manner is using dedicated hardware infrastructure. Also if you have a need to integrate H.320 ISDN into the call then hardware is pretty much your only option (yes - there are still areas of the world doing ISDN video conferencing!).
What would be your recommendation for a video network deployment? Hardware or software? Is there a cost difference?
Polycom’s perspective is that hybrid deployments will be the way to go for many organisations, certainly in APAC. It would make sense for organisations to deploy much of their conferencing infrastructure using software solutions, but they should still consider their specific requirements for the bridging component.
For greenfield customers looking to deploy a mainly software client video solution, connecting desktops and tablets then SVC and software bridging will most likely be the most cost-effective and flexible solution. Now with new solutions such as Polycom RealPresence One that customer now has the option of paying a subscription for their video collaboration infrastructure rather than having to purchase the relevant licences outright. Of course, purchasing perpetual software licences is still an option.
For existing customers who have room systems deployed, or for a greenfield customer considering a mix of desktop and room systems, then a hybrid approach makes much more sense. They could deploy the entire infrastructure, except the bridge, as software only, and then use a hardware conference bridge for the video processing. Of course they could have a mix of bridging within the network, and have both hardware and software bridges. This is ideal if there is a wide mix of conference requirements, ranging from large volumes of small-scale conferences from software devices through to large town hall meetings connecting multiple conferences.
When you look at the overall total cost of ownership between a hardware and a software solution there is unlikely to be much difference once server costs, virtualisation software etc. are taken into consideration. An organisation with existing virtualisation infrastructure and spare hardware capacity might find cost benefits in a software approach as they would already have many key components. The key difference though with software is the flexibility it delivers in how licences and infrastructure components can be deployed.
Polycom’s perspective is that it is important that customers have all the options available to them when considering their video infrastructure. Hence we have ensured that all our infrastructure components are available as appliances for simple rack-&-stack deployment as well as software virtual machines. In addition by taking advantage of software and offering a subscription via Polycom RealPresence One customers now have the greatest range of options, not just from a deployment perspective but also from a purchasing perspective.