|
 |
|||
|
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
We hear of VoIP as the wave of the future, and a more cost effective means of providing both voice and data linking within the organization over a single network. A careful examination of all details usually suggests that there are costs involved in upgrading the existing data network, misunderstanding in what is involved, and exaggerated claims. This article is designed to address some of the fundamental steps required to implement and thereby help determine the value of a VoIP conversion internally. This article does NOT address a VoIP phone carrier who provides voice lines into your buildings nor does it discuss the merits of VoIP vs. tranditional carriers. |
|
|
|
|
|
Article |
|||
|
We hear of Voice over Internet Protocol, VoIP, as the "new and improved" way to route your voice and date traffic on a single internal network. Other claimed advantages are extension portability (moving an IP phone to a new Ethernet outlet without any wire changes), external access via public Internet connections, and lower costs of operation as well as lower long distance charges. Before you jump into this, it is wise to understand the concept of VoIP, integrating VoIP with the data network, and to discuss what is involved. VoIP may be appropriate, then again it may not be. Traditionally Voice was handled on its own internal network. A common phone box, often called a Key System or a PBX was the central controller. The Phone Company lines plugged into the central controller. Each internal extension was connected to the central controller with a dedicated pair of wires. The central controller was often sufficiently sophisticated to provide voice mail, call back, camp on / call back, and even route Phone Company DID (direct in dial numbers) to a specific extension. This was a voice only network (although many times dial up modems and similar devices used this system for lightweight data transfers). The voice transmission was very good and highly reliable. In effect, within the internal network, a dedicated pair of wires were allocated from the hand set to the destination (another internal hand set, a pager, an outside line, or other devices – and the central controller provided full attention assuring that it completed the circuit and provided a wired dedicated pair of wires to the circuit to assure phone performance. Nothing was shared. When the extension was unused, the line was unused. Telephone technicians understood the switching and were well trained in the specific central controller. Data networks differ from voice networks. Data flows in discrete packets of data over a shared line. The data packets are routed via network gear to its destination. A given data transmission (e.g. an e-mail) consists of dividing the data in to numbered packets, dispatching the packets (often over differing routes) and reassembling the packets at the destination. If a data packet is lost, the receiving hardware requests a resending of the lost packet. The transmission is not complete or noticed at the receiving end until all packets are received and reassembled. There is no requirement to finish a transmission within a specified time frame. If the data demand exceeds the line's available transmission capacity (bandwidth), data packets waited in line, like a traffic back up on a highway, until it could get sent. If the back up is severe, packet loss happens and lost packets ultimately are resent. This was and is acceptable for data. Putting voice on a data network raises many issues. The voice is converted to data, the data packets are sent, received, and reassembled. Lost packets are resent before reassembly. This has to be done fast enough for the voice conversation to be acceptable. Choppy, staggered, intermittent voice communication is not acceptable and results when voice packets are not received in time. Voice packets compete with data packets for bandwidth. If you have more bandwidth than demand, things usually work well. If data transfer needs arise during the voice conversation, and bandwidth demand exceeds capacity, data packets back up or are lost. This can severely degrade voice and result in unacceptable voice transmission. To provide acceptable voice transmission, the following needs to be addressed on the data network:
The traditional phone service person is not usually sufficiently trained or skilled in network management, priority setting, and use allocation as it is more sophisticated. Troubleshooting of the data network usually requires a different type of resource. Adding voice to the data network may require more data trained technical people to assure the system performs. This will add to the network cost. If you have multiple locations, you need to link them. Data network links may be via a Virtual Private Network over the Internet and both locations may be well served via a DSL service to the internet. Running voice over this is likely to be a disappointment or a disaster because the bandwidth may be too small and because the priority coding of voice packets are removed by the ISP and priority codings are not followed on the Internet. This negates the priority assignments – and usually results in unacceptable voice transmission. Often a substantially more costly point to point dedicated leased line will be required. This added cost needs to be factored in. What are the reasons to implement the change? If the existing phone system is performing well and does not merit a replacement due to age, reliability, obsolescence, or other factors is it wise to upgrade? Often, you can interface a VoIP system at one location to a traditional one at another if you wish. What about long distance charges? Remember that your phone system, whether it be VoIP or conventional, connects to the Public Switched Telephone System (PSTN). The PSTN charges you for long distance calls – and they do not have different rates for connecting the PSTN to a VoIP or a traditional internal phone system. This means that there is no long distance savings with internal VoIP systems. If you have an empty building and have to provide data and voice, VoIP may be the best choice for that location - nothing is being retired early and everything needs to be procured. Often these cases strongly favor VoIP. Prior to dedicating funds to the VoIP project, allocate funds for a network engineering review to determine exactly what is needed to assure VoIP success. This engineering review or assessment is much like an engineering study for construction where the engineering needs are determined and specific upgrades are detailed and costs are forecast. You would not build a structure without this, why try to put your voice communications together without such a prudent technical review? Be sure you understand what voice quality the proposal will deliver. Skimping on network capability may produce voice quality that is extremely inferior or, worse yet, totally unacceptable. Beefing up the network after install is only more costly. As with everything it is cheaper to do it right the first time. If you cannot afford to do it right the first time, the project probably does not merit approval. As with all other business investments, carefully review the project's need and the costs and understand the benefits before committing resource to implement VoIP. Perhaps VoIP would be of merit in your organization. If you are interested in discussing this further, feel free to call for a no cost, no obligation initial discussion. Gene Bohn |
|||
