In the first half of 2013, SIP trunking spiked 23% and has increased by 6 fold by 2017 according to Infonetics Research. Organizations are starting to move to all IP based communication solutions. In terms of UC (Unified Communications) and VC (Video Conferencing) adoption has been lagging, however VoIP (Voice over IP) has been the primary driver for the technology shift to IP. There are critical components that need to be addressed when an organization is migrating to an enterprise VoIP architecture. The migration project should be split into phases and it makes sense to dissect the deployment map into three components: the “customer-premises network”, the “inter-office network” and the “voice trunks”. Unlike the customer-premises network, businesses don’t control the inter-office network and the voice trunks, however, they have choices and they can choose one service provider over another, based on the value of their service offerings.
1. For customer-premises networks, SIP is the future. Any enterprise VoIP equipment that has no SIP support will have potential problems in terms of future compatibility. Yes, you can install SBCs to normalize and therefore accomplish interoperability if you already have a legacy equipment install base, however, if you have the choice, SIP support is a must. Another point to consider is hosted vs on premise PBX solutions. As their names suggest, on premise PBX is a physical equipment that installs at the customer premise, ideally with built-in HA (High Availability) and/or with a hot-failover capability (tip: ask your PBX vendor if they can keep the calls alive during the failure of the primary PBX). If you have the option, choosing two different locations for the PBXs will add additional protection for any catastrophic failures. In the hosted PBX scenario, the IP-PBX provider will take over all that maintenance functions, however, in this case, all the uptime concerns will put pressure on the WAN link connecting the offices to the hosted PBX and that brings us to the 2nd design element – the voice trunks.
2. Both with hosted PBX and on-premises PBX, the backend connection of the PBX will connect the phone to the party on the other side of the line. Usually your hosted PBX either can natively connect to a phone trunk (SIP trunk, PRI trunk, or analog POTS), or will use a WAN connectivity technology to connect to the phone trunk at the service provider location. The voice trunk provider will then handle converting the call traffic into the PSTN (Public Switched Telephone Network) so that inbound and outbound calls can be completed. If you are planning to keep some PRI or POTS analog trunk channels, you should be aware of the associated costs, as these older technologies are not as cost effective as the more modern SIP trunks. At the same time, if you opt in for a SIP trunk, it is critical to have a WAN connection to the SIP trunk that is reliable and is high quality as these factors will directly affect the voice quality. This brings us to the last and arguably the most important design element – WAN reliability and quality.
3. For a successful VoIP implementation, it is critical to have high reliability and high quality WAN links that create the inter-office connectivity, as well as the connectivity from the headquarters to the voice trunks. In the earlier analog days, each branch office maintained its own voice trunk, i.e. PRI or POTs that connects to the local PSTN, which is the analog channel that connects the office phones to the telco network. With these legacy technologies, every office required the associated equipment and therefore costly maintenance. With VoIP-based solutions, the PBX and the connectivity to voice trunks can be moved to a central location, i.e. the regional headquarters. Branch offices can then connect and make use of that centralized equipment via the IP tunnel between the branch offices and the headquarters.
There are several options for inter-office connectivity, each with their own pros and cons. MPLS networks can provide good quality, but at a high cost, and can be limiting for organizations with diverse locations. MPLS, by definition, is managed by a single carrier and therefore the offices that are not in the service area are either out of luck or will suffer from “inter-ISP coordination syndrome”.
Cost-effective WAN services on the other hand (such as ADSL, cable and fiber) have been improving in terms of quality and in most cases offer an excellent price point for enterprise VoIP implementations. However, they usually lack appropriate SLAs (Service Level Agreements) on QoS (Quality of Service) and can suffer from fluctuations in link performance.
In recent years, a new option emerged for inter-office connectivity that combines the best of both worlds, i.e. a tightly controlled QoS on cost-effective bonded WAN links. This is possible by virtualizing two or more WAN links at the branch locations, thereby creating a better performing WAN link at a fraction of the cost of MPLS. The idea is to Broadband Bond WAN technologies (such as DSL, cable, fiber, etc.) that can intelligently inherit the best parts of those broadband links and therefore provide an MPLS-like network for a fraction of the cost.
Given the wide range of options and various migration paths, moving to an all IP-based communication architecture should be carefully planned, tested, and rolled out. The first step should be replacing the inter-office WAN architecture so that they all connect to a centralized or hosted PBX as this provides the biggest bang for the buck with the highest cost savings without changing the voice trunks. Once the WAN virtualization is successfully completed, migrating the PRI trunks to SIP trunks can be implemented in phases.
Cahit Akin, CEO, Mushroom Networks, Inc.
Mushroom Networks is the provider of SD-WAN (Software Defined WAN) and NFV solutions capable of Broadband Bonding that enables self-healing WAN networks that route around network problems such as latency, jitter and packet loss.