SIP Trunking Architecture

VoIP comes in many flavors, configurations and set ups. Many aspects have changed since the early days of VoIP. In the beginning, many VoIP implementations were ad hoc and pretty simple. They lacked many of the features we’re used to now but for the most part, they worked.

VoIP deployments have come a long way since then, especially in business environments. Enterprise VoIP solutions today are feature rich and extremely inexpensive to boot. In addition there are many people & companies with proven expertise and track records in the VoIP space who can provide services for businesses of any size and type.

Hosted VoIP and SIP Trunking

Businesses going for VoIP generally have two options – either go for hosted VoIP services or set up their own on premise solution. On the face of it, these are clear cut options. However there are nuances involved in these alternatives as well. For instance companies can choose to purchase everything from a provider – including even hardware like desk phones or headsets. Or they can purchase only certain services and get their own hardware.

On the side of the SIP trunking architecture there are several options as well. Enterprises can do everything on their own – right from purchasing equipment to configuring the IP-PBX boxes. Or they can chose to do some parts on their own while purchasing other services from external contractors. VoIP is inherently flexible which means companies have a lot of options when it comes to upgrading to VoIP.

Types of SIP Trunking Architecture

On premise deployments can be configured in a variety of ways to suit the needs of the company. However there are two main types of architectures which are commonly used – centralized and distributed solutions. Both types have their own pros and cons but there is no such thing as the ‘best’ alternative. Which solution is better will depend on many factors including:

• Availability of resources like money, time and expertise
• The needs and requirements of the business
• Criticality of functions and features

Centralized SIP Trunking Architecture

When SIP implementations are deployed in centralized topologies, the trunks are installed in central data centers. When a call is made from an employee to another, it is usually referred to as an internal call. These calls are routed through the corporate data network which doesn’t incur any charges. When external calls are made, they are all routed through the central data center and then handed off to the service provider as a whole.

Advantages & disadvantages

Certain aspects become immediately obvious with regards to centralized topologies. There are economies of scale which can result in dramatic savings. For instance all the supporting infrastructure for SIP trunks need only be installed in a single location. Maintenance, routine repairs and troubleshooting become easier and quicker. If something goes wrong, it is an easy task to pinpoint the problem since all the equipment is in a single location.

However these benefits also come with other issues. Having all SIP trunks at one place also means that the business has a single point of failure. Should something go wrong with the SIP service, the company will lose the ability to make and receive calls completely. Depending on the problem, it may even take several hours to rectify. Imagine the havoc it can cause if your phone system is unreachable by customers or partners. It can result in loss of brand image, revenues and even contracts.

Additionally centralized topologies require much more planning and advance preparation for increasing capacity. For instance the total capacity of the trunks should be able to support the peak demand across all sites. If the estimated numbers are wrong by even a few percentage points, it can result in overwhelming the network or degrading the quality.

Distributed SIP Architectures

In distributed topologies, the planning and preparation is much easier. In the vast majority of cases, the business need only replace existing analog lines with SIP trunks. For the most part, it is a 1:1 replacement. Each office location will have its own SIP trunks, independent of other sites. So some locations may have fewer or more trunks than others, depending on the traffic.

Advantages & Disadvantages

In this case there is no single point of failure for the organization to worry about. If the trunks or equipment at one location has a problem, calls from other employees are not affected. The business as a whole does not shut down. It makes it easier to plan for business continuity or disaster management planning.

The organization need not invest in all the equipment at once, instead opting to do so in a phased manner throughout. If demand from a particular office increases or decreases, sessions can easily and quickly be added. Problems at one site do not affect others, so troubleshooting need not take away precious time from the enterprise as a whole.

On the other hand, distributed topologies do not gain the financial benefits of centralized architectures. SIP trunks have to be individually configured, repaired and maintained. It is more intensive in terms of personnel and financial commitments. Another drawback is that extra session capacity cannot be pooled i.e each trunk has to support peak demand at each location. This may necessitate purchasing more capacity which lies unused for the most part.

As we can see, both types of implementation have their own considerations to keep in mind. While there isn’t a universal best option, companies can even gradually switch from one type to another over the years as their requirements change significantly.