This article gives you a brief walkthrough of how to protect your network from being overloaded by voice and video traffic using call admission control in Microsoft Lync Server 2010 communications software.
Author:
Rui Maximo with Subbu Chandrasekaran and Thomas Binder
Publication date: June 2010
Product version: Microsoft Lync Server 2010 communications software
One of the big concerns that IT often raises when they consider deploying VoIP and video conferencing is whether their IP network is prepared to sustain the network load imposed by this media traffic. VoIP and video conferencing require high bandwidth and low latency to offer a good end-user experience. The concern is legitimate. While upgrading existing IP networks may be the ideal solution, the reality of this option is likely to be cost-prohibitive and often occurs over a long period of time. In the interim, what other options are available?
Call admission control provides a way to prevent degradation of quality by limiting the number of concurrent calls over limited bandwidth links such as wide area networks (WAN).
Lync Server 2010 supports highly adaptable audio and video codecs that can adjust to varying network capacity. The RTAudio and RTVideo codecs make it possible to maintain good media quality in degraded network conditions. However, to prevent degradation of audio and video quality that users can perceive, Lync Server introduces support for call admission control. It is now possible to prevent users from establishing calls that would result in quality degradations for everyone.
Call admission control offers more flexible control for IT professionals to architect their network traffic. This will help prevent unexpected spikes in calls from affecting the entire network, line of business applications, and impacting the quality of current calls. Call admission control protects the network and prevents Lync Server traffic from consuming all the bandwidth available on the network. Unlike other call admission control solutions that are available from different vendors, call admission control in Lync Server does not require additional hardware; it is built into Lync Server and Microsoft Lync 2010.
When the network link is about to become oversubscribed, call admission control in Lync Server 2010 re-routes new calls through the PSTN or offloads the media portion of the call over the Internet. This offers more flexibility for handling call spikes instead of just letting the oversubscribed sessions fail. When re-routing calls, call admission control attempts to redirect the audio through the Internet first before attempting to re-route it through the PSTN.
Call admission control provides reports on blocked calls and redirected calls. Similar to the PSTN, blocked calls is preferable to poor quality calls. Redirecting calls to alternate routes (Internet or PSTN) is even better than blocking calls. This data helps organizations fine-tune their call admission control policy for their particular network as they go. This helps reduce the anxiety of optimizing call admission control at first use.
Figure 1 illustrates a call being authorized through the WAN link. Figure 2 illustrates the same call; however, instead of being blocked, the call is re-routed through the PSTN. Figure 3 illustrates the same call. In this case, the media is redirected over the Internet. Notice that the signaling path is still through the WAN link.
Figure 1. Call established over WAN link
.jpg)
The following steps describe the process of establishing a call between the calling party (Alice) and the called party (Bob) over the WAN link.
|
1
|
Alice initiates call to Bob.
|
|
2
|
Bob’s Lync 2010 receives a call notification.
|
|
3
|
Bob’s Lync 2010 checks the call admission control policy to determine whether the call can be established.
|
|
4
|
Bob’s Lync 2010 accepts the call.
|
|
5a
|
Call is established and audio flows across the WAN link.
|
Figure 2 shows the call media being redirected through the Internet.
Figure 2. Call media redirected over the Internet
.jpg)
The following steps describe the process shown in Figure 2 of the call’s audio being redirected over the Internet.
|
1
|
Alice initiates call to Bob.
|
|
2
|
Bob’s Lync 2010 receives a call notification.
|
|
3
|
Bob’s Lync 2010 checks the call admission control policy to determine whether the call can be established.
|
|
4
|
Call is not allowed to be established over the WAN.
|
|
5b
|
Call is accepted and directs audio path to be redirected to the Internet.
|
Figure 3 illustrates call admission control re-routing the call through the PSTN.
Figure 3. Call re-routed to PSTN
.jpg)
The following steps describe how the call is rejected over the WAN link, but is re-routed to the PSTN.
|
1
|
Alice initiates call to Bob.
|
|
2
|
Bob’s Lync 2010 receives a call notification.
|
|
3
|
Bob’s Lync 2010 checks the call admission control policy to determine whether the call can be established.
|
|
4
|
Call is not allowed to be established over the WAN.
|
|
5b
|
Call cannot be routed over the Internet.
|
|
5c
|
Call is re-routed to PSTN.
|
|
6c
|
Call audio flows across the PSTN.
|
Let’s explore how call admission control works in Lync Server. This article will provide you with a high-level understanding of how call admission control works in Lync Server.
Call admission control is managed by the Bandwidth Policy service in Lync Server. This service is automatically installed as part of every pool and benefits from the high availability design of the Enterprise pool. Bandwidth Policy services across multiple pools within an organization automatically synchronize with each other to maintain integrity of bandwidth subscription. You can configure the Bandwidth Policy service by using Windows PowerShell or Microsoft Lync Server 2010 Control Panel.
Before you configure call admission control, it is important to understand the terms, network regions, network sites, and network links used in call admission control.
Network regions consist of multiple network sites. A network region represents a network hub or backbone. Network regions are interconnected through a wide area network (WAN) link.
Network sites identify locations within a network region. A network site represents a physical location that belongs to an organization such as a branch office or a regional office.
Network links refer to the WAN link that connects two network regions. Such WAN links have limited bandwidth capacity compared to LAN links; therefore, call admission control is enforced on such network links.
Configuring call admission control involves the following steps:
-
Identifying network regions.
-
Identifying network sites.
-
Determining the IP subnets assigned to each network site.
-
Identifying the network links that connect network regions. For each network link:
-
Determine the maximum bandwidth capacity.
-
Define the bandwidth capacity (meaning, the call admission control policy).
-
Determining the network route between every pair of network regions.
It is likely that the networking group in your IT organization will already have defined your organization’s network regions and network sites and identified the IP subnets for each network site and the bandwidth capacity of the WAN links between network regions (steps 1-4a in the previous list). The hardest part will be chasing down this information.
After you obtain this information, you can decide how much bandwidth capacity to allocate for audio and video across each network link and determine the preferred route between network regions.
Call admission control policy defines the following information:
-
Maximum total bandwidth to allocate for audio
-
Maximum total bandwidth to allocate for video
-
Maximum bandwidth that can be allocated for a single audio call (session)
-
Maximum bandwidth that can be allocated for a single video call (session)
To optimize per-session bandwidth use (see 3 and 4 in the previous list), you need to consider the type of codecs being used to avoid under-allocating sufficient bandwidth needed for a particular codec or over-allocating bandwidth by setting the maximum bandwidth per session too high. In a subsequent NextHop article, we’ll provide you guidance on how to optimize your call admission control policies based on the codecs allowed.
Lync Server doesn’t enforce call admission control at layer-2 networking, which is managed by routers. To ensure that network traffic is routed in the network path that you want when multiple paths are possible, routing will need to be enforced at the router level.
Call admission control is enforced at the client level. It is enforced by the receiving client not the call initiating client as explained earlier in the “Design” section of this article. Because only Lync 2010 is call admission control aware, clients such as Communicator 2007 R2 and Communicator 2007 won’t enforce call admission control when those clients are the recipients.
Call admission control ensures quality of experience of audio and video calls by enforcing that audio and video traffic managed by Lync Server 2010 doesn’t take more bandwidth than configured by the policy. The support of call admission control in Lync Server 2010 makes Enterprise Voice a robust VoIP solution for the enterprise. Call admission control helps IT professionals proactively prevent introducing additional high bandwidth, low-latency traffic into saturated network links, which would quickly degrade the IP network under peak-traffic conditions. When combined with layer 2 Quality of Service (QoS) at the routing level, call admission control helps ensure high quality audio and video between endpoints across WAN links. Leverage call admission control to your advantage!
Communications Server Resources