Increasing numbers of enterprises are deploying unified communication and collaboration (UCC) applications and services. This is not surprising since IP-based UCC brings definite advantages—enhancing productivity, boosting responsiveness and reducing total cost of ownership (TCO). Since UCC applications are designed to run on an organization’s data network, IT personnel should be prepared to deal with the effects.
Vendors should work with IT in order to determine which applications the company requires, how they will be implemented and what impact they will have. In an enterprise environment, a solid network infrastructure is required in order to build a successful UCC system. UCC applications place strict requirements on IP packet loss, packet delay, and delay variation (jitter). Network administrators need to activate the quality-of-service (QoS) mechanisms throughout the network to ensure that the network infrastructure is high-performing, resilient, secure and complies with open standards. Resellers should also understand how technologies like voice over IP (VoIP) and videoconferencing impact bandwidth. Here is a look at two common UCC technologies and the affect they have on data networks.
VoIP traffic is much more sensitive to latency, jitter, and packet loss than most network applications. Vendors should help IT Implement QoS before the organization implements VoIP. To implement end-to-end QoS, voice traffic must be given higher priority than data traffic on every link. By implementing QoS traffic prioritization, VoIP will provide the best call quality. In addition, companies should route voice traffic into its own virtual local area networks (VLANs) to separate the voice traffic from the data traffic. VLANs will put the IP phones on their own broadcast domains and IP subnets. Vendors should help IT find a troubleshooting tool for VoIP traffic. Network troubleshooting tools developed primarily for VoIP are available from several companies. Finally, vendors and IT managers should conduct an analysis in order to determine potential network bottlenecks and develop a baseline:
Video application deployments are on the rise. Organizations that have the most success with their multimedia deployments know to put ample resources into their initial network design and to take into account building monitoring and management capabilities in order to prevent problems. As explained in this paper from Jisc, videoconferencing is about interaction, a two-way exchange of information. Video streaming is a related technology, but is generally used for one-way transmission rather than interaction. For effective interaction, participants in a videoconference must be able to communicate in real time. A high level of processing inside of the videoconferencing endpoints occurs in order to reduce latency. However, there is only so much the endpoints can do to deal with traffic/packet loss or delays across the networks. The Jisc paper goes on to outline the usual metrics used in order to determine how a network is performing:
- Latency (End-to-End Delay) – With the codecs currently used in videoconferencing endpoints, a latency of 100 ms can be expected in both the coding and decoding processes; however, 300 ms is a reasonable maximum target for getting audio and video through the endpoints and network.
- Packet Loss – For videoconferencing, delivery of real-time traffic across networks has tended toward the use of UDP (User Datagram Protocol), which is an “unreliable” delivery protocol. The network will send and receive packets without considering what, if anything, has been delivered. It is up to the application to decide whether or not any processing is required in order to counter the effects of missing or delayed packets.
- IPDV (Inter-Packet Delay Variation)/Jitter – In a network that is experiencing congestion, the delay between packets can vary considerably. Usually, the IPDV of packets across a network will be measured in milliseconds; it should be noted, though, that when things go wrong, the IPDV can be measured in seconds. Across the network, IPDV will normally be well below 20 ms with only a few milliseconds being the norm.
- Bandwidth – Bandwidth usage can vary considerably. Care must be taken when specifying nominal versus peak bandwidth requirements for videoconferencing traffic, and allowances should be made at those points in the network where bandwidth may be limited—for example, by QoS policing—in order to ensure that the peak traffic level can be processed correctly.
Data networks must be planned and managed while keeping in mind the increasing performance demands of users and applications, especially UCC. Flexibility and scalability are key requirements for ensuring the ability to make future capacity increases that expanding businesses require. Resellers that can communicate these requirements effectively will win the trust—and the business—of organizations considering VoIP and videoconferencing tools.
Please feel free to comment on this article below.