Small businesses converging voice, video and data traffic sometimes end up with haunted networks, but you can exorcise the demons. Affordable QoS mechanisms are available to oust the goblins that cause delay, jitter, bit-rate errors and dropped packets.
QoS tools use prioritization to improve a network’s ability to deliver predictable service in support of specific applications and types of traffic. Traffic prioritization ensures that real-time applications, such as VoIP and video teleconferencing, get adequate service for optimal performance and other data-intensive applications aren’t left in limbo.
An investment in QoS can improve the speed and quality of such applications, and reduce infrastructure and transmission costs — an added benefit for businesses with their eyes on the bottom line.
Make your network performance howl
Although the capabilities and quality of IP solutions have advanced and their costs have decreased, the solutions are often plagued by noise and service interruptions caused by delays, jitter and dropped packets. The root cause, though, is that the core networks are ill-equipped to handle the traffic.
There are three levels of QoS tools that can remedy these problems:
* Best-effort service provides basic connectivity without prioritization of specific flows or applications, so there is no guarantee of service. A good example of this is first-in, first-out queuing, in which packets are stored when the network is congested and forwarded in the order they arrived once traffic eases.
* Differentiated service identifies and classifies traffic to assign priority and meet latency requirements within available bandwidth. This results in faster handling, more predictable service and fewer dropped packets. The network administrator predetermines the priority of specific types of network protocols.
* Guaranteed service reserves network resources for specific types of traffic. The network administrator develops bandwidth classifications, and the bandwidth is divvied up among them to guarantee availability to the highest-priority traffic while serving lower-priority traffic on a best-effort basis. This prevents low-priority flows from superseding high-priority flows.
Different QoS tools are available to enable better network service to applications, and should be selected based on the type of traffic being managed:
* Classification tools. These identify and in some cases, mark packets so they can be properly managed by other QoS tools.
* Congestion avoidance tools. These prevent queues from filling and subsequently dropping packets.
* Congestion management tools. These manage the queue and flow of network traffic when bursts — high levels of voice, video and data traffic — occur simultaneously.
* Link-efficiency tools. These fragment large data packets and interleave real-time packets between the fragmented packets so they can be sent without delay or interference.
* Shaping/policing tools. These limit the amount of bandwidth the packets can use, to pace traffic at a controllable rate and prevent potential queue overflow.
All QoS implementations require a network-monitoring system to track traffic, identify performance degradation, and regulate service to give precedence to voice and video traffic. Networks require switches that support QoS and 802.1p standards — a standard that accelerates the classification and filtering of traffic.
Trick or treat: What’s in QoS for whom?
There are multiple benefits to implementing QoS. The quality of voice and video transmissions improves and, on the back end, QoS allows administrators to control network resources and monitor traffic to ensure critical applications receive appropriate service.
All of that, of course, makes it less daunting to adopt convergence technologies which can deliver ghoulish savings:
* Users at different sites can make extension-to-extension calls over the existing data network rather than requiring additional land lines.
* Businesses can implement 64-party audioconferencing bridges that save hundreds of dollars a month on conference-call services
* Session Initiation Protocol trunks can cut phone bills by as much as half by connecting the phone system to traditional telephone lines via the Internet over a data E1 line, reducing the use of voice E1 lines and associated costs.
What’s more, with QoS, teleworkers can use a IP phone from their home office over a VPN link, making it possible for them to handle calls as if they were in the corporate office, increasing their accessibility. “Softphones” — software on laptops — with QoS also are an excellent option for employees who travel frequently, because they allow remote users to call extensions at the main office for free via their computer.
That’s not to say that QoS is without challenges. QoS tools can relieve network congestion, but sometimes there is just too much traffic for the given bandwidth. To ensure quality, bandwidth must be high enough to support latency of 150 milliseconds or less.
Switch failure is also a concern for businesses that use IP phones with dual-port switches to connect the phone and a computer through a single drop, applying QoS to route voice traffic ahead of data traffic. If the two-port switch fails, the phone and the computer both lose connectivity. Likewise, users will lose connectivity if the LAN switch fails. Recovery systems must be in place to get the network back up and running as quickly as possible. Many businesses operate a POTS (plain old telephone service) line to maintain some access during these times, and some businesses avoid the potential challenge altogether by implementing a LAN for each type of device, though this option is more costly.
Queue overflow — when the number of incoming packets exceeds the rate at which the network can process them — can also be a challenge because a full queue cannot receive any additional packets and will drop those that arrive last, regardless of their priority level. This condition usually happens when voice packets are not prioritized ahead of data packets or QoS is not implemented correctly, causing jitter and echo. Network administrators should implement congestion avoidance tools, such as virtual LANs, so voice packets receive greater priority and the queue does not fill up. They also should establish mechanisms that drop low-priority packets first.
Finally, here are a few suggestions about getting the most out of your network with QoS:
* Determine if QoS is right for you by weighing the benefits to the network applications against the cost of implementing QoS.
* Assess network traffic to determine which applications are most critical to the business and which require the most bandwidth and availability. Failure to conduct a network assessment is the most common mistake and can lead to significant expenditures to correct configurations.
* Select a vendor that provides training and guidance for the implementation process, as well as a reputable managed-services provider if the business does not have an in-house IT professional.
* Establish minimum requirements for: bandwidth among multiple office sites; availability of service for priority applications; traffic-management performance — typically a latency of 150 milliseconds or less; and the user experience and satisfaction with the performance of the phone system — there should be crystal-clear sound without jitters, drops or echo.
* Assess the results to verify that the service is performing at the desired level.
QoS is at the leading edge of network management, and is necessary for the optimum performance of networked applications. Remember, however, that managing QoS is an ongoing process because application and network demands change constantly. If you install QoS tools to optimize network infrastructure and improve the performance of bandwidth- and data-intensive applications, you’ll have to actively manage the complete system; but the network and those it supports will be the better for it.