Over the past eight years, Kevin Pashuk has overseen three major videoconferencing projects in Ontario: building a telemedicine network for a medical school from scratch, upgrading a system for a college and overhauling one for a private school.
So he knows that if there’s a business need for it, IT managers shouldn’t fear the network demands of videoconferencing.
“Decisions are made quicker,” says Pashuk, now chief information office of Appleby College in Oakville, Ont. “Video adds an element of the building of community, the building of network, the ability to build trust across space and time.”
Unfortunately, not everyone sees it that way. System manufacturers and industry analysts admit there is still handwringing among a number of IT and network managers when told to prepare for on-premise videoconferencing.
“I believe there is great fear, uncertainty and doubt on the part of network managers because they haven’t looked into the situation carefully,” says Andrew Davis, senior analyst and partner at Wainhouse Research, which specializes in unified communications.
Worries about bandwidth, capacity, interoperability between systems and the lack of a globally accepted directory system for reaching outside the firewall are chief concerns.
“It is for the most part the most bandwidth-intensive and latency sensitive application a network manager is going to face,” warns Zeus Kerravala, senior vice-president of research at the Yankee Group. “VoIP doesn’t even come close.”
“For people who have deployed real-time communications like voice over IP networks, video is nothing new,” says Ian Gallagher, a unified communications product specialist at Cisco Systems Canada. “The fact is that higher bandwidth is almost a moot point.”
The main work, he says, is making sure end points can handle the additional bandwidth needed, and then properly configuring call control to restrict the number of video calls that can be made over restricted bandwidth.
On the other hand, he admits it can be a big step for those who have never deployed any form of quality of service or voice over a WAN depending on the kind of videoconferencing needed.
Use in business is rising
Videoconferencing has been around in one form or another for 15 years. But its use in organizations has risen steadily in the past three years due to several factors: The increasing shift from proprietary protocols to SIP (session initiation protocol), a steady drop in the price of bandwidth, improved image quality and the cutting of travel budgets thanks to the recession.
The past 12 months has seen an acceleration of moves in video by traditional unified communications equipment or software manufactures and entries by newcomers. For example, earlier this year Cisco bought Tandberg SA to give broaden its range into mid-sized systems, a move countered by Hewlett-Packard Co., which has concentrated on its high-end hosted Halo systems, by forming a partnership with Vidyo Inc.
“Video can be a demanding beast,” explained Darren Podrabsky, “which is why we invested in Vidyo.” The company is leveraging the H.264 Scalable Video Coding (SVC) compression codec to shrink bandwidth. However, the standard isn’t widespread.
Meanwhile, Polycom Inc. has formed partnerships with Avaya Inc., Siemens Enterprise Communications Group and Juniper Networks; and Logitec Int’l. bought LifeSize Communications Inc.
On this side of the border, startup Magor Communications Corp. of Ottawa is trying to find a foothold with three-, two- and one-screen mid-range systems.
Does size matter?
Planners have to think of the pros and cons of the systems. So-called telepresence systems, whose goal is to reproduce an almost life-size image of the person on the other side of the screen using High Definition (HD) need tailor-made rooms for best effect.
But they eat bandwidth. For example, at a demonstration for IT World Canada, a high-end Tandberg T3 system with a trio of 65-inch screens swallowed 12 Megabits per second of HD video at 60 frames a second. But dropping the frame rate and running at standard definition on a desktop can squeeze a call down to 384 kbps.
Desktop systems, ranging from PC monitors to deskphones with small screens, have mixed acceptance – unless you have a decent-sized monitor, why look at someone’s tiny visage?
On the other hand, desktop systems are often linked to Microsoft Office (and the coming Lync) or Lotus Sametime clients and don’t need the setup and scheduling that bigger systems do, making ad hoc calls easier.
Room-based systems, with one or two large monitors, are considered an answer for teams. However, Yankee Group’s Kerravala calls them “a solution looking for a problem.” Often too many people appear on-screen, negating the intimacy good video can bring. Similarly, a view of people sitting at a table isn’t inspiring.
And typically, room-based systems aren’t equipped with the best sound systems, a big hindrance to communications.
Experts weigh in
So how can an organization get the best use of a VC system? We asked a range of experts.
Companies that are the most successful with videoconferencing recognize their network and applications teams have to work together to properly support any real-time application, says John Bartlett, an Acton, Mass.-based member of voice and video consulting firm NetForecast Inc.
Amir Hameed, who leads a systems engineering leader at Avaya Canada that advises customers, warns organizations against creating separate voice and video strategies.
Jayanth Angl of London, Ont.-based Info-Tech Research, advises network managers to analyze their current traffic to ensure it can deliver video with adequate quality of service. Latency, jitter and packet loss and sources of congestion will impact performance, he said.
Setting different QoS policies in routers and switches for branches and types of endpoints is essential, said Cisco’s Gallagher, as well as setting admission controls to ensure video doesn’t impact other applications. If bandwidth is limited, call controls also need to be created to keep the system from being overused.
Bartlett also warned that setting QoS packet markers for a local area network isn’t the same as a WAN, where the provider owns the network. You’ll have to make sure your LAN techniques map to the WAN — and that will get complicated with multiple providers.
Videoconferencing between suppliers or customers is another headache, one that can be solved either by setting up a dedicated network, or going over the Internet through a session border controller.
Over the years Kevin Pashak has overseen a wide range of problems. At the Northern Ontario School of Medicine, a joint venture between two universities some 600 km apart, the videoconferencing network was shared between the institutions and Orion, a province-wide university research network. Under those circumstances a good service level agreement is key, he says.
“If you’re going to do videoconferencing,” he adds, “engage someone who understands the back end. Otherwise you’ll get into a whole lot of grief.”
Another lesson he came away with is the need to set up an automated management system so booking a video room online is simple.
He’s now CIO at Appleby College, a senior high school advanced enough in technology that students get new laptops every two years but where the VC system was in its infancy. So while Pashuk crafts a five-year plan to expand the video system, his staff has been making interim moves, like videoconferencing an outside lecture through Skype on a laptop connected to a projector.
His plan envisions that any room in the institution where learning takes place will have the ability to host a videoconference. To get there, the school’s Internet capacity as been more than doubled, new firewall, core and edge switches are in place and a new video bridge is in the works.
“If you remove geography, space and time as barriers,” he says, “the dreamers get really excited.”
So, hopefully will the network staff.