Getting high quality VoIP for your company

VOICE-OVER-INTERNET PROTOCOL has taken off because it helps organisations reduce costs, and service providers are now attracted by revenue potential. Operators can use VoIP to rapidly deploy new value-added and high-margin applications and services. To gain the quality and reliability businesses need, networks have to support high-quality voice and video, protect it against threats and misuse, and assure performance and availability.

In designing a VoIP solution, service providers need to consider how their choices will address latency, jitter, bandwidth, packet loss, reliability, and security.

Minimising latency

Latency — the time it takes a packet to make its way through the network to the destination — does not necessarily degrade sound quality but can cause hesitations during voice conversations.

Miller
End-to-end latency should be less than 150ms for toll quality phone calls. Jitter and voice quality Jitter — the time difference between when a packet is expected to arrive to when it actually arrives, is usually caused by dynamic changes in network traffic loads or packet re-routing.

Most media gateways have play-out buffers but these cannot eliminate severe jitter and this can cause voice quality issues.

Bandwidth requirements

Allocating bandwidth between services in a voice and data network requires careful consideration.

Allocating too little bandwidth to voice services could result in unacceptable quality.

Consider also, impact of bandwidth priority for sensitive VoIP packets; trade-off between compression and voice quality, and projected peak use.

Changes in call initiation rate, voice encoding method, packet creation rate, compression and silence suppression can all change bandwidth requirements. Compensating for packet loss During network congestion, routers and switches can overflow their queue buffers and be forced to discard packets.

Packet loss is undesirable for applications such as web browsers and file transfers but not critical because the protocols they use, can retransmit. Real-time applications are less tolerant.

Some voice packet loss can be tolerated if it is spread out over a large amount of users and voice quality is not generally affected if the packet loss is less than five percent of the total number of calls.

Ensuring reliability

Failover strategies and redundant links between network devices, media gateways and media gateway controllers are needed for rare network failures.

Media gateways and controllers that actively detect the status of their next-hop address decrease likelihood of a large service disruption.

Connecting media gateways direct to the router can enable network devices to immediately detect link failure and take appropriate action.

Setting security

VoIP networks are vulnerable to many of the same security risks as data networks, including Denial of Service (DoS) attacks, service theft, tampering, and fraud.

The security device must support VoIP protocols like SIP, MGCP, and H.323, and associate state at the signalling layer with packet flows at the media layer.

Session border controllers can filter VoIP sessions and screen traffic classification and packet processing engines to protect equipment in the network against DoS attacks.

To dynamically open and close ports for VoIP traffic, firewalls need to understand VoIP signalling protocols.

Operators should establish firewall policies to protect communications between servers and VoIP end-devices and to restrict VoIP communication based on authorised end-devices or traffic sourced or destined for a particular IP address or interface.

Firewalls can also be used to segment a VoIP network to ensure appropriate priority and policies are applied.

Operators can deploy intrusion prevention systems to help detect and prevent certain attacks, such as manipulated Dynamic Host Configuration Protocol (DHCP) messages or flooded forwarding tables.

Traffic-engineered MPLS

Multi-Protocol Label Switching (MPLS) is the most modern network functionality for VoIP and other real-time applications.

Traffic Engineered MPLS technologies are designed from the start to address the complexities and high availability requirements of carrier-grade VoIP and other premium services.

Fast re-route, auto-bandwidth provisioning, and virtual path traffic engineering (TE) tunnels guarantee the high level quality and reliability we expect from telephony services.

Choosing a vendor

Companies should assess VoIP equipment vendors according to their:

* Ability to support different network transport service models. Care must be taken in addressing VoIP’s extremely rigid QoS (quality of service) and reliability requirements. Choosing an experienced vendor with expertise in deploying large, complex multiservice IP networks will pay significant dividends;
* Commitment to open standards in products. Any vendor should be actively developing voice strategies that interoperate with all VoIP protocols. Without this, VoIP systems could become as proprietary as legacy voice systems;
* Ability to support multiple protocols to create a VoIP solution that can handle future system migrations and incorporate products that support a different protocol; without requiring wholesale infrastructure redeployments or significant network upgrades; and,
* End-to-end support for all VoIP protocols. Vendors must provide solutions that work in both single and multi-protocol environments. (Andy Miller is vice-president of Juniper Networks Asia Pacific’s enterprise division.)

Andy Miller is vice-president of Juniper Networks Asia Pacific’s enterprise division

source news : star-techcentral.com