Microsoft Lync Server 2010 : Speak to Your Branch Offices
Some of the topology changes in Microsoft Lync Server 2010 make it easier to deploy the full range of unified communication technologies.
Alan Maddison
Among the challenges you face as an IT administrator are a lack of dedicated support and persistent budget constraints. This is exacerbated when you’re charged with managing not only your own central office, but also a handful of branch offices.
Consider some of the common technical constraints often associated with supporting branch offices: concerns about network reliability, bandwidth usage and end-user support. Consequently, when planning for and deploying unified communications (UC), many companies restrict certain aspects of UC functionality from branch-office users. While the reason may be to preserve bandwidth and save costs, it leaves those remote users less connected and available.
Of the core technologies embodied in UC, the most likely to be left out of remote deployments due to cost, convenience and support is voice functionality. With Lync Server 2010, Microsoft has made a number of enhancements that make these decisions and restrictions worth revisiting.
Lync Server 2010 now offers significantly improved support for branch offices, with significant changes in topology design, changes in how clients and servers interact, and the introduction of key technologies such as Call Administration Control (CAC) and Survivable Branch Appliances (SBAs). These upgrades address the common challenges associated with deploying UC to branch offices.
Speak Up
Survivability is one of the fundamental requirements for any organization looking to deploy enterprise voice capabilities to a branch office. Survivability refers to the ability of a branch office to place and receive calls within the site and be able to place and receive Public Service Telephone Network (PSTN) calls.
Even with the prevalence of cell phones, nearly all organizations require remote offices to have reliable PSTN connectivity. This may be to ensure the company can still conduct business or so employees can contact emergency services if needed. Nevertheless, the PSTN access requirement remains a challenge for IT teams looking to deploy enterprise voice to remote offices.
With Lync Server 2010, Microsoft worked with a number of partners (Network Equipment Technologies Inc., AudioCodes Ltd., Dialogic Inc., Hewlett-Packard Co. and Ferrari Electronics AG) to deliver a hardware-based solution that can ensure survivability for branch offices. An SBA provides voice resiliency in a single chassis. This appliance is loaded with Windows Server 2008 R2, the Lync Registrar service, the Mediation Server role and a PSTN gateway. For larger branch offices, you also have the option of deploying a Survivable Branch Server. This is a standard Windows Server 2008 R2 server with the required Lync components and a PCI ISDN card.
Because an SBA is intended to run at remote sites, Microsoft and its hardware partners designed them for easy deployment, seamless resiliency and remote management. Besides ensuring voice resiliency for branch-office users, SBAs can also support fault tolerance and local high availability.
SBAs have features like redundant hot-swappable power supplies and seamless failover between SBAs. These features are all designed to minimize management requirements, maximize uptime and ensure a consistent and positive user experience.
Once users are connected to the SBA in the event of a wide area network (WAN) failure, not only will it maintain voice survivability through the PSTN or the Internet, it will maintain resiliency in the event that domain controllers become unavailable, plus it can let users perform certificate-based authentication.
Always On
Once you’ve ensured survivability for branch-office voice capabilities, the next consideration is the WAN connection. There’s the basic question of how much bandwidth is available. Also, there are three major factors you must consider when evaluating a network connection’s overall performance: latency, jitter and packet loss:
- Latency is a measure of how long it takes a packet to traverse the network; latency is often caused by serialization, propagation or queuing
- Jitter is a measure of the differences in latency values for individual packets
- Packet loss is a measure of how many packets are dropped at some point during transmission
Fortunately, when a WAN is oversubscribed or experiences high packet loss, high latency or jitter, there are many mitigation techniques. These include retransmission, forward error connection, queuing and buffering at the network, OS and application layers. These techniques can often offset the negative effects of oversubscription. Therefore, many applications are tolerant of slow or inconsistent response times.
Even Lync has a certain tolerance for constrained bandwidth and packet loss, which is more remarkable when you consider that Lync also has some of the most highly adaptive codecs (specifically RT Audio and RT Video) on the market today. Under the heaviest loads, though, even the most-efficient codecs—those that are specifically designed to optimize a user’s voice experience in a changing network environment—may be insufficient.
These factors are the underlying reason that the International Telecommunication Union specified round-trip latency for a voice-over-IP (VoIP) call should be 300 ms or less. When you consider that VoIP has additional latency overhead caused by the encoding and decoding of voice packets, it becomes clear that providing voice capabilities to branch offices requires careful planning.
Control the Call
CAC is a critical technology for enterprise voice planning. CAC helps provide an efficient and effective solution for managing bandwidth. It also helps ensure that users receive a consistent experience in terms of voice quality.
By helping you ensure that quality doesn’t degrade, CAC ensures that enterprise voice-call quality remains consistent over time. This helps improve the user’s experience and minimizes the number of support calls generated by quality issues.
You manage CAC through policies. There are separate policies for audio and video, which you apply to users and enforce through the Bandwidth Policy Service. You can restrict the amount of bandwidth used by Lync clients for both audio and video. You can also restrict the number of sessions, or concurrent calls. This helps ensure that each session or call has a minimum amount of allocated bandwidth. You can take that one step further and let certain users bypass the policy enforcement for their location.
CAC is essential for resolving some of the challenges associated with WAN connections and call quality. Nevertheless, you can’t overlook another key technology that has enhanced support in Lync Server 2010. Quality of Service (QoS) is a tried-and-tested solution for networking technology that helps prioritize traffic in a network environment. In earlier versions of Lync—including Office Communications Server 2007 R2—you could only assign a Differentiated Services Code Point, or DSCP, value to the actual media traffic.
Lync Server 2010 supports DSCP marking of Session Initiation Protocol (SIP) signaling traffic. This ensures that call traffic is appropriately prioritized. Lync also supports DSCP marking for Application Sharing and File Transfer traffic, which allows for a greater degree of granular control.
In Case of Emergency
Once organizations start deploying VoIP beyond their centralized location, it’s critical they don’t overlook the affect on emergency service calls. This is particularly important when you have branch offices or users routing voice traffic through a centralized datacenter that may be in a different part of the country.
In North America, Lync Server 2010 now supports Enhanced 911 (E911). This means it can use Location Services to provide accurate location information to emergency service personnel. Based on the National Emergency Number Association i2 Architecture for Enhanced 9-1-1 Services, the Location Information service is a Web Service that’s part of every Lync front-end server.
You can determine location by two different methods, either manually or automatically. The manual method involves letting the user enter street addresses into the Lync client. You could also automatically define central locations through categories like mac addresses, subnet ranges, switches, ports and wireless access points. You can also use this location information as part of the user’s presence information.
Lync Server 2010 has made tremendous enhancements, including comprehensive support for branch offices. The functionality provided by SBAs, CAC, QoS and E911 add a high degree of power and flexibility. However, incorporating these technologies into your UC topology has tremendous implications from a design and planning perspective.
When deploying voice to branch offices, you need to design and configure a number of elements, including bandwidth policy profiles, network regions, links between regions, routes between regions and network sites. In larger organizations or companies with numerous sites, you must carefully consider these requirements in order to ensure that your topology design is both effective and efficient. Add in QoS planning and location services support, and you begin to understand the significance and enormity of this undertaking.
Fortunately, the end result of a company-wide Lync Server 2010 deployment is significant ROI. In addition, there are a number of qualitative improvements due to enhanced collaboration and participation.
.jpg "Alan Maddison")
Alan Maddison* is a 16-year veteran of the IT industry, during which time he has focused primarily on Microsoft technologies. For the last five years, he has worked as a consultant focusing on delivering professional services. He can be reached at alan.maddison@microsoft.com.*