For the past 20 years, the videoconferencing industry has been dominated by systems designed for conference rooms, board rooms, and other shared facilities. Attention is now shifting to personal systems that can be deployed cost-effectively in corporate and home offices as well as in cubicles and contact centers. With the shift from ISDN to IP video networks now largely completed, many enterprises are increasingly aware of the possibilities for increased efficiencies that are possible when visual communications are integrated into the existing workflow.
One of the key enablers in today's integration efforts is Unified Communications (UC), a concept in which a single user interface is used to access a variety of communications modalities--presence, instant messaging, voice, and audio/video/web conferencing--on an ad hoc basis. The UC interface can be a separate application, such as a buddy list application, or the UC capabilities can be embedded into other business applications such as CRM or ERP systems, contact centers, team workspaces, internally and externally facing websites, and custom dashboards and mashups.
Mindshare in the emerging UC marketplace is dominated by Microsoft, according to several surveys conducted by Wainhouse Research; and with the introduction of Microsoft Lync, Microsoft continues to define and redefine the baseline capabilities to which other UC solutions must measure up. Microsoft Lync 2010 provides a single interface based on identity and presence that unites voice communications, IM, and audio, video, and web conferencing into a rich, contextual offering. With Lync, Microsoft has set its eyes clearly on the transition to IP telephony and voice as a network application, and on rich media communications. Microsoft has also created the ability to offer these enterprise conferencing capabilities both as CPE deployments and as hosted services.
Based on interviews with large and small enterprise users of videoconferencing, we are aware that many customers are looking to enhance small deployments of room systems with large deployments of personal/desktop systems while looking to Microsoft, and Lync specifically, to provide the needed infrastructure support and ease-of-use. Much like audio conferencing has migrated from the conference room with a speakerphone to the personal desk phone with meet-me ad hoc bridging, the future of videoconferencing is linked to the enterprise desktop and the requisite infrastructure components (presence and bridging) for ad hoc video meetings. Microsoft has not missed this vision, and video, while secondary to telephony in the Microsoft marketing campaign, is one of Lync's attention grabbers.
Lync's video capabilities are the sole focus of this note. For those enterprises looking to Lync for the foundation of their video deployments, we have identified solutions or strategies that fall into several distinct categories--including offerings from Microsoft and those from third-party vendors.
Microsoft Lync Video
Videoconferencing is integrated into Lync 2010, Microsoft Lync's desktop client (formerly the Communicator client corresponding to the former OCS server, which is now simply called the Lync server). The integration makes scheduling easy and enables click-to-call dialing. Lync supports an ad-hoc conferencing model in which users can start with an IM text message and escalate the session as desired to a voice or video call. Ad hoc communications sessions and functional escalation are some of the key differences between UC-based videoconferencing and legacy videoconferencing applications.
Lync Server 2010 supports both point-to-point (P2P) video calls and multipoint calls (audio and/or video); multipoint calls require Microsoft's AV Conferencing Server which is built into Lync. Lync provides support for 720p video and multipoint calling using a "voice activated switching" (VAS) algorithm in which the image of the active speaker is displayed. This contrasts with the more popular and processing-intensive "continuous presence" (CP) display in which the images of multiple participants are displayed simultaneously. Continuous presence video images with Lync require a third-party multipoint control unit (MCU).
Figure 1: Microsoft Lync P2P or VAS video call (top); Microsoft OCS/Lync with CP video via a third party MCU (bottom)
According to the Lync specifications document, HD video (1270 x 720) and VGA video (640 x 480) are supported for peer-to-peer calls between users running Lync 2010 on high-end computers; HD video in Lync 2010 requires a quad core processor for encoding the video and a dual-core processor for video decoding. (Frame rate has been bumped up to 30 fps vs. 25 fps with the previous-generation OCS.) The resolution viewed by each participant in a single conversation may differ, depending on the video capabilities of each user's respective hardware.
Microsoft has developed its own audio and video compression algorithms, and these are the "favored" codecs in point-to-point calls between Lync 2010 clients. Microsoft's Real Time Audio (RTA) delivers wideband audio (7kHz) while Microsoft's Real Time Video (RTV) supports high definition 720p resolution. Neither RTA nor RTV are ITU Recommendations and are hence considered "proprietary." A distinguishing feature of both RTA and RTV is a forward error correction algorithm that provides some level of quality sustainability in the presence of IP network packet loss. When communicating with an industry-standard SIP system, Lync supports H.263 video with CIF resolution and G.722 wideband audio. Lync can connect to an H.323 system only through a third-party MCU or gateway. Note that most systems today from leading videoconferencing vendors support both H.323 and SIP, although not necessarily Microsoft’s version of SIP, in which case a protocol transcoding gateway would be required.
While nothing has been formally announced, many informal signs and "leaks" suggest that the next version of Lync, dubbed Wave-15, and probably two years away, will add an H.264 scalable video codec (SVC) to the mix, perhaps becoming the preferred video algorithm for Lync-based videoconferencing. Polycom, for one, has announced that it is working on an SVC implementation for Lync and that it may be one of several SVC options Microsoft may support. At the present time, several vendors, including Vidyo and Radvision, offer SVC-based videoconferencing solutions, however none of them, including the anticipated Polycom implementation, are interoperable, and none of them work with Lync as of May 2011.
Direct Endpoint Integration
From a technical perspective, in order for an endpoint to register to Lync it must support Microsoft's version of SIP; NTLM (NT LAN Manager, a suite of Microsoft security protocols); ICE and TURN for NAT firewall traversal; Conference Control Channel Protocol (CCCP) for signaling; SRTP (Secure Real-Time Protocol) and TLS (Transport Layer Security) for media encryption; and H.263 for video compatibility. This combination allows an industry-standard endpoint to interoperate with a Lync endpoint. Polycom recently announced general availability of its direct integration with Lync using Microsoft's RTV codec in Polycom’s HDX video endpoints and RMX MCUs. Polycom and LifeSize both support H.263 video integration. (LifeSize has taken a step further by providing automatic SIP settings discovery and registration, making direct integration automatic for the administrator.)
With direct endpoint integration, a third-party peripheral such as an IP telephone or an IP videoconferencing system will appear to the Lync Server 2010 as a standard Lync endpoint. From a user's perspective, the direct integration approach has several advantages. Generally, the user doesn’t have to do anything at all. To authenticate a device, a person simply uses his Lync user name and password. (This would be true for the user behind a Lync gateway as well--a Lync user name and password are all that is required.) Perhaps more importantly, there is usually no additional cost for the endpoint hardware, although some vendors may charge if a software upgrade is required to integrate with Lync.
Directly registering devices to Lync allows administrators to use Lync's bandwidth management, directory management, and NAT/firewall traversal capabilities. Direct integration also eliminates the obvious costs and single point of failure associated with a separate Lync gateway (see our discussion on Lync gateways below).
Figure 2: Polycom HDX with direct integration to Lync. Note the multiple (continuous presence) images versus a single image one would get from Lync alone.
Direct endpoint integration implies that the endpoint should support all Microsoft features and functions, which often, but not always, includes seeing the buddy list on the videoconferencing or IP phone device. Polycom has announced direct endpoint integration with RTA support for its CX series of IP phones and for the CX5000 USB-based Unified Conference Station (formerly known as RoundTable); the CX5000 also supports the RTV video codec.
One disadvantage of the direct Lync endpoint integration approach is that it generally takes time for a vendor to do the software integration, and the integration has to be tested across multiple hardware devices and software revisions, including new releases of Microsoft Lync itself, etc. Typically, only the latest products in a vendor's videoconferencing portfolio can support the direct registration mode. Direct integration often means compromise as well--lower quality audio and video algorithms (720p maximum in Lync verses 1080p in most of the newer video endpoints), lack of content sharing, etc, although these restrictions might disappear with the next generation of videoconferencing systems should they support Microsoft's screen sharing protocol. Registering video endpoints directly to the Lync server is best suited for small video deployments where costly investments in video infrastructure are to be avoided. Direct integration is unlikely to be attractive to those who are moving to Lync but have a large installed base of video systems and video infrastructure (gateways, gatekeepers, NAT/firewall traversal). In these situations, a gateway approach (see below) may be the optimum solution.
Direct MCU Integration
Very similar to direct endpoint integration, direct MCU integration involves having the enterprise MCU support the Microsoft codecs and protocols. In this case, the MCU itself will appear in the Lync buddy list. The most common use case for this approach is the "meet-me bridge" whereby users call into a conference hosted on the MCU.
The advantage here is that the bridge can connect callers using Lync desktop video clients with others using H.323, SIP, and/or PSTN and ISDN devices. And each user can call into the bridge using his own familiar user interface (web, hand held remote, tablet, control panel, etc.) The disadvantage is that if the enterprise does not already own a bridge, considerable expense may be involved, and scaling to support hundreds or thousands of users may be prohibitively costly.
Video Gateways
A video gateway is a device that sits between an island of Lync users and an island of non-Lync users (typically H.323 or SIP room and desktop systems) and provides the connectivity and transcoding services needed to enable interconnectivity. Unlike an MCU which is programmed to handle multipoint conferencing connections, a gateway is a "pass through" device with logic specific to one-to-one connections. One important gateway specification is the number of simultaneous one-to-one connections the device can support.
A Lync-compatible gateway will extend the reach of Lync to the world of group videoconferencing, preserving Lync's single interface experience while seamlessly integrating the vast market of deployed video conferencing devices into the Lync environment.
Functions provided by a Lync gateway should allow any H.323 device or resource (MCU or ISDN gateway) to be displayed in the Lync contact list, and users could search Lync for any legacy videoconferencing endpoints, which could then be added to their contact list. In principle, users should be able to make a call to any legacy video device exactly the same way they would call another Lync user. This would include support for SIP forking, which enables inbound calls to simultaneously ring multiple devices.
Figure 3: Integration of the Radvision UC gateway to the Lync environment
A transcoding gateway obviously needs to support all the algorithms and protocols of all islands of connectivity. The advantages of the gateway approach are that the vendor has only once device to support, so that features and upgrades can be released much faster; interesting features can be added more easily, such as H.264 HD support for Microsoft clients; presence information for legacy systems; and most importantly--support for a wide variety of devices, including MCUs, and mobile phones. With a gateway, Lync users can connect to new and old industry standard desktop video systems, room systems, and even telepresence suites.
The disadvantage lies in cost, scalability, and single points of failure. Some of the vendors' gateway solutions involve multiple, expensive boxes and the number of simultaneous gateway calls is limited. (In all fairness, enterprises that have an existing video deployment and are moving to Lync are likely to have some of these infrastructure pieces already installed). Some customers may opt for a signaling gateway only (this means that the same audio and video codecs must be used on both sides); and some may opt for a signaling/media transcoding gateway to support full HD video interoperability with Lync. Like the direct integration approach, gateway vendors will need to keep up with software revisions released by Microsoft.
Software Plug-Ins
With prior versions of Microsoft's UC platform, some videoconferencing vendors took an approach that used the APIs Microsoft exposed in the Office Communicator client to integrate H.264 video systems. In this scenario, the vendor's own desktop video client software provided the video communication capability.
With this technique, the plug-in exposed a drop down menu or a tab at the bottom of the OCS interface which was then used to launch a separate videoconferencing application. Efforts were made to have the user interface look as integrated as possible, but this was done through on-screen design rather than true software integration, which was not possible because of the limited APIs available in Office Communicator. This solution required varying degrees of R&D on the part of Microsoft's third-party partners, depending on the extent of the integrated functionality and the user interface alignment.
Microsoft modified the plug-in capability with Lync, which requires third party partners to adapt their integrations to the API now available in the Lync client. An example is Cisco’s release of Cisco UC Integration for Microsoft Lync (CUCIML) soon after Lync was released. The Cisco integration exposes a variety of Cisco UC-based capabilities from Lync despite the loss of the tab interface.
Other Solutions
Unified Communications, and Lync in particular, will be used by customers for other types of visual communications besides videoconferencing. VBrick, for example, has introduced the concept of Video Buddies, implemented by the company's streaming gateway for OCS/Lync. With this solution, the presence engine can link to video content organized by "channels" under VBrick’s media management system, and by clicking on a "channel", video is streamed to the user's Lync client, a very useful solution for education and corporate training. For collaborative applications, Lync provides the engine for real-time collaboration using voice and text chat where video assets rather than documents may be the "data" being discussed--an application of interest to military and medical users particularly. Other vendors have built solutions to capture Lync sessions and prepare them for webcasting/streaming.
Approaches to Microsoft Lync Videoconferencing
A number of vendors provide integration solutions with Microsoft Lync. In the table below, we list different vendor integration solutions, noting that at the time this article was written, only Lifesize has earned qualification from Microsoft.
Accordent: Accordent provides an automated, end-to-end infrastructure for creating, managing, delivering and measuring the effectiveness of video-based content. The company's "Accordent Media Management System" (AMMS) software can be integrated with Microsoft Lync, giving users anywhere in an organization the ability to capture presentations and conference sessions from their desktop and publish them to the AMMS. Once captured, this content can be approved, transcoded, published, and distributed over the network with secure distribution assured through user authentication and multilevel permission features.
Avistar: Avistar is offering "Avistar C3 United for Microsoft Lync" which is part of the company’s C3 product line. Like many third-party products in the Lync space, C3 United is a newer version of the company’s OCS-compatible product. The Avistar solution is a right-mouse-enabled, customizable plug-in for the Lync client as well as a companion conferencing server that preserves the robust server features of Lync but overcomes several deficiencies Avistar perceives in Lync's videoconferencing capabilities.
The basic user interface consists of right-clicking on a user’s name and then left-clicking on "Place Avistar C3United Video Call" in the pop-up menu. One of the unique features of the Avistar solution is that it can bring Lync into the virtual desktop integration (VDI) environments favored by many large enterprises. Avistar also supports adding people "on the fly" to an ongoing video call.
Other key features of the overall solution include continuous presence multipoint calling, telephony features like call hold and forward, full interoperability with H.323 systems, forward error correction, comprehensive bandwidth management and call admission control functions, and NAT/firewall traversal support.
Cisco: Via Cisco's acquisition of Tandberg, Cisco offers a very comprehensive (but relatively expensive for greenfield situations) solution for those who want to bring their videoconferencing deployment--including desktops, rooms, and telepresence suites--into the Lync world.
The heart of the Cisco solution is the Tandberg Video Communications Server (VCS), a platform that connects H.323 environments to SIP environments. The VCS is both an H.323 gatekeeper and a SIP back-to-back user agent and does SIP-to-H.323 internetworking. The VCS is also a call policy server that defines and enforces bandwidth rules for a network. Special features specific to Cisco are a videoconferencing FindMe function similar to a PBX-based hunt group and an OCS or Lync relay that can make any standards-based video device registered to VCS appear in the Lync or OCS contact list. FindMe leverages the simultaneous ring functionality of Lync, a feature that enables an incoming call to ring several extensions either in parallel or sequentially. (The first extension that answers establishes the session.)
VCS also enables secure firewall traversal. The VCS establishes a SIP trunk to the Lync server and publishes the presence of H.323 and SIP endpoints (registered to the VCS) to Lync. Cisco/Tandberg endpoints (or other industry standard videoconferencing systems) can then communicate to Lync users via H.263 with CIF or SIF resolution at 15 frames per second. (Cisco/Tandberg endpoints support Microsoft’s versions of SRTP, ICE, and SIP).
A second Cisco device (optional), the OCS or Lync Advanced Media Gateway, adds transcoding between RTV and H.264 and supports up to 720p resolution as well as RTV VGA. The gateway also supports calling into an MCU. In an unrelated product dimension, Cisco also sells a PrecisionHD USB camera.
LifeSize: LifeSize has an uncertified build that works with OCS and Lync but does not support RTV. Hence their endpoints can register to the Microsoft server but cannot connect with the Microsoft MCU and cannot support H.264 when in a Lync call, only H.263. LifeSize has since licensed RTV and is working on a Lync gateway via a development partnership with Microsoft. The company is striving to have full Lync certification by the end of 2011. The LifeSize Lync gateway will transcode between H.264 and RTV as well as between "regular" SIP and Microsoft SIP. LifeSize is also a member of the Unified Communications Interoperability Forum (UCIF) and, through this organization, will be in a position to support scalable video coding when and if such an algorithm is embedded in Wave-15.
Figure 4: LifeSize / Microsoft OCS (left) LifeSize Team System with embedded MCU (right)
Polycom: Polycom is taking the direct endpoint integration path and, at the present time, appears to be the only company doing so. While the videoconferencing industry's history is littered with failed, fragile, or meaningless partnerships, we believe Polycom is attempting to position Microsoft as #1 in its list of seven strategic partnerships as well as positioning Polycom as one of Microsoft’s most strategic partners and certainly the Redmond giant's favored video company.
According to Polycom, "the Polycom HDX series telepresence systems and Polycom UC Intelligent Core Infrastructure are the industry's first telepresence solution designed to be fully interoperable with Microsoft Lync." What this means is that Polycom's HDX codecs and RMX bridges now have native support for Microsoft's RTV (but not RTA) algorithms. (HDX units will connect to Lync using the G.722 audio codec, which Lync supports.)
In addition, Polycom offers the CX series of IP phones that are optimized for Lync and natively run Microsoft Lync 2010 Phone Edition. The Polycom CX series includes the CX5000, a USB camera device introduced to the market years ago as Microsoft Roundtable. Polycom has also implemented Microsoft’s Centralized Conference Control Protocol (CCCP) in its HDX units, making it possible for them to seamlessly participate in multipoint video conferences hosted on Lync's audio/video server.
Polycom claims that its product development team is working closely with Microsoft engineers to develop future Lync products such as SVC (mentioned previously). Microsoft and Polycom are founding members of the Unified Communications Interoperability Forum, and word has it that this forum is working on an interoperable SVC video algorithm. Polycom competitors are quick to point out that they too have a "partnership" with Microsoft or at least a press release to that effect.
Radvision: Radvision's technology for connecting the company’s videoconferencing products is based on a signaling and transcoding gateway. Radvision began working with Microsoft unified communications products several years ago, early in the life of LCS and OCS.
In October 2010, Radvision announced that it had entered into a strategic partnership with Microsoft to deliver advanced integration to Microsoft's UC solutions--a combination of hardware and software integrated with Lync that will be offered through joint marketing and sales initiatives. The focus would be the ability to bring Lync's UC advantages to standards-based (read legacy or installed systems) videoconferencing devices as well as to new mobile solutions. In November, 2010 Radvision announced the SCOPIA UC Gateway for Microsoft Lync 2010.
The Radvision relationship gives Microsoft a nice balance in videoconferencing partners, with Polycom taking one approach and Radvision another, thereby providing customers with ample choice. And unlike the situation in years past, both partners now offer videoconferencing appliances and software endpoints as well as infrastructure solutions. The earlier strategic relationship between Microsoft and Tandberg is clearly dead, given Tandberg's acquisition by Cisco.
VBrick: VBrick is shipping the "OCS Streaming Gateway" which sits between their IP video platform and the OCS/Lync infrastructure. The Gateway registers VBrick streaming video channels as virtual users with presence (Program Title). They are also Communications Enabled Business Process Agents (CEBPA) that can enable third-party applications to push messaging and initiate video playback on one or many desktops via OCS/Lync. Applications include emergency broadcast alerts with one-to-many streaming video that can reach thousands of desktops concurrently via multicast or unicast with zone-based intelligent distribution.
The VBrick Gateway extends the Microsoft UC deployment to include additional applications such as TV distribution to the desktop, live training channels, surveillance with real-time collaboration, and one-to-many "executive town hall" broadcasts that retain the OCS/Lync context.
Figure 5: VBrick/OCS
Vidyo: In 2009 Vidyo announced a plug-in for OCS 2007 R2. This plug-in allows users to launch a Vidyo videoconference from their Microsoft Office Communicator client. We believe Vidyo is working on a plug-in approach which will allow Lync 2010 clients to interface with users who are on the VidyoDesktop client.
Concluding Thoughts
There is clearly no "one size fits all" when it comes to enabling videoconferencing services with a Microsoft Lync deployment. Different approaches will make sense to different customers, largely depending on their current videoconferencing deployments. For example, those with a large installed base of endpoints and infrastructure may find the UC gateway to be the most efficient approach to connecting to Lync users. For Lync users who want to add room video systems, deploying products that support direct integration may be optimum.
No matter what solutions are considered today, the options are likely to change over the next few years as Microsoft upgrades the native capabilities of Lync itself while the videoconferencing vendors continue to differentiate themselves by providing higher levels of performance, appliance options that are independent of the desktop PC hardware, and infrastructure products to bridge with the legacy world.
Andrew W. Davis is Senior Analyst & Founding Partner, Wainhouse Research, and E. Brent Kelly, Sr. is Analyst & Partner at Wainhouse Research.
