QoS is a a Quality of Service product that allows you to identify different traffic types within each connection to the Pacific Wireless network so that each traffic type can be given its own bandwidth guarantee and priority. Subsets of your total bandwidth can be allocated, for example, to VoIP, terminal services, VPN and Internet and then prioritised in accordance with your own specific requirements.

Why have it?

QoS is required when bandwidth is a limited resource (which is usually the case when buying services from a telecommunications provider) and the network administrator wishes to:

	Enable limited resources to be shared evenly or fairly according to the importance of the traffic
	Ensure that time sensitive traffic (e.g. VoIP, terminal services) is passed through the network without delay

Both requirements can be met by implementing the two techniques available within our QoS product:

	Bandwidth Guarantees (minimum or maximum)
	Prioritisation

Bandwidth Guarantees allow you to allocate subsets of the available bandwidth to different traffic types. For instance, you might allocate 35% to Voice, 35% to Prority Data and 30% to all other traffic. By setting a minimum guarantee, each traffic type can burst above its allocation whenever the other types are not using theirs. Alternatively, a maximum guarantee can be set to prevent those traffic types with associated usage costs (for example Internet or Priority Queues) from exceeding a preset level.

Prioritisation allows you to minimise network delays and loss for time-sensitive and critical data packets that must arrive at their destination in the shortest possible time. Bandwidth guarantees alone may not eliminate delays that can occur when the total link bandwidth is approached or when network nfrastructure is unusually busy. For prioritisation to be fully effective, every network device through which your data passes must forward your priority packets ahead of all low priority packets, including those of other customers on the network.

Why is it the right choice?

	Minimum or maximum bandwidth guarantees can be specified for each data service (Internet, VPN, VLAN)
	Within each data service, you can further classify your traffic into Priority Voice, Priority Data and Standard
	Traffic classes can be user-defined using IP address and IP ports
	Each traffic class can have its own bandwidth guarantee within the parent data service’s own guarantee
	Enabling prioritisation means Priority Voice goes first, then Priority Data, then Standard
	On-net prioritised packets go ahead of all Standard packets, whether yours or those of other customers

How do I get it?

Your sites need to be connected to the Pacific Wireless network with either Wireless Connect or DSL Connect, however some limitations may apply when used with DSL Connect. For wholesale partners, QoS is also available on Carrier Connect and on connections to Powertel and Nextep. For multi-service Wireless Connect links, bandwidth guarantees can be specified for each data service without having to define traffic classes, enabling the link bandwidth to be divided between Internet, VPN and VLAN.

How does it work?

As soon as data packets enter the Pacific Wireless network, they are inspected and matched against the customer-specified QoS criteria for both bandwidth guarantees and prioritisation. If a packet matches any of the criteria, it is marked in such a way that all Pacific Wireless equipment that will process it on its journey can enforce the QoS options. If a packet does not match the QoS criteria, it is also marked to ensure that it cannot gain unauthorised access to network-wide prioritisation.

On a multi-customer network, end-to-end time-sensitive QoS can only be fully effective if both the customer and the Service Provider identifies and prioritises their traffic within their allocated bandwidth, and the Service Provider can also prioritise that traffic above non-priority traffic belonging to other customers.

A key feature of Pacific Wireless on-net QoS is its ability to do just that. The following example illustrates why this is important:

Customer 1 has a three site VPN with a Head Office site (HO) and two remote sites (RS1 and RS2). Traffic between the remote sites and head office traverses a link between two network points of presence (POPs). Customer 1 has enabled QoS for the VPN so that priority voice data always takes precedence over other traffic within the VPN. Customer 2 has a two site VPN. They conduct large file transfers between the sites that are not time-critical. If the Service Provider only prioritises Customer 1’s traffic within the VPN Router and not at the Network Switch, Customer 1’s priority voice may still be subject to delays if either the Network Switches or the link between them gets busy as a result of the activity of multiple other customer networks such as that of Customer 2. Ask your service provider to explain the scope of their prioritisation scheme.