Frame Relay is a data transmission protocol. It gets its name because it places data, and the information required to transmit it properly, into a “frame” that is made up of several fields. It places the data to be transmitted into a specific field, surrounded by four other fields, which contain information about where the data frame should go and how it should be handled. A user installs dedicated circuits to connect its locations to the Frame Relay network of their carrier. Then, the carrier utilizes what's known as “Permanent Virtual Circuits” or “PVCs” to transmit the data within the network.
The main advantage of Frame Relay over point-to-point is that it enables you to only need one connection from each site to enable you to be linked to all sites. For example, if I have 4 offices that all need to be connected to each other (called a “meshed” network”), I would need 6 circuits. But, with frame relay, I need only 4. Frame Relay enables me to do that job with 1/3 fewer circuits. Plus a user can also connect a PVC to the Internet in order to incorporate BOTH its LAN/WAN networking and its Internet Bandwidth needs into a single network architecture.
When buying Frame Relay, the user will select the bandwidth of the PVCs that it purchases. This is the “fixed” data speed, or bandwidth, that they have. The user can also select a “burstable” bandwidth / data rate. Burstable means that on an as-needed basis you can “burst” to a higher speed, and only pay for that additional bandwidth when you actually use it.
One example of this is that a client has a normal ongoing need for 512 kbps of bandwidth among its locations. Plus, once a month it backs-up all data from each site to the main site, which requires a “burst” to 1.544 mbps for a few hours to accomplish the data transfers. In a point-to-point service, the client would be forced to buy 1.544 mbps bandwidth to all locations, even though it would only be fully utilized once each month. For this and similar reasons, Frame Relay has been the predominant type of data communications service for business for many years.
So what does a client need?
It depends upon a number of factors, but mostly ....
(1) how many locations do they need to network, and
(2) what types of communications do they want to use the network for.
Today, very few firms are looking at implementing Frame Relay unless they already have it in place. New installations are typically choosing between MPLS and Carrier Ethernet options. I will not seek to get into detail on these other two services here, but suffice it to say that if a client only has two or three sites they may be fine with point-to-point, or just connecting via Internet and using Internet-based applications to accomplish their voice and data communications. If they have more locations, signficiant bandwidth needs, and/or a desire to perform voice, data, and video communications among the sites, they are most likely better off with MPLS, mostly because MPLS will enable them to control quality of service, which is most critcal when utilizing mixed services, and real-time services such as voice and video.
In laymens terms ..... I'll try to simplify even further.
1. With a point to point connection you have paid for a connection between two points. You pay a committed rate for that connection per year, or per month. There are several variables that affect the cost.
A) How far apart are the two points
B) Do multiple telecommunications carriers connect the points or just one carrier
C) Are the points in different Cities, States, Countries, Continents
D) How fast do you want the connection to be
E) Do you want the connection always available or only available when you request it (e.g. ISDN or even dial-up --- they are both versions of point to point, just tempoary)
2. With Frame Relay, or MPLS or ATM or any other number of shared connection systems .... you are buying only the "last mile" of your connection on both ends, while you are sharing the middle with other users. You pay your network carrier to segregate the traffic inside the middle (typically described as a "cloud" because it's an abstract concept, your network provider can implement it any way they need to) and to provide you guaranteed bandwidth. Some carriers will sell you the ability to "burst" your speed higher so that you can occasionally (but not always) use higher speeds on demand. Your connection speed is limited by the lower of either the "last mile" connection speed on either end or by the committed rate you bought from the carrier. The carrier will use a name like CIR, Committed Information Rate to describe the rate you have purchased.
A) Distance not really a determiner of price
B) Do multiple telecommunications carriers connect the points or just one carrier (you still have to pay last mile charges on both ends)
C) Are the points in different Cities, States, Countries, Continents (could affect tariffed rates and taxation, plus regulatory)
D) How fast do you want the connection to be
E) Usually you are paying for an always on connection, called a permanent virtual circuit. You really only consume network resources when you send data but it's always there for you immediately.
3. Finally you can stop paying for that part in the middle completely by using Virtual Private Networks across the Internet. In this case you still have to pay for access on both sides, but nobody pays for the middle. There are of course some tradeoffs:
A) Since nobody is guaranteeing the middle, you are on your own to support it and troubleshoot it
B) Your traffic will comingle with public internet traffic. It's up to you to ensure it is secure. You do this by employing a well regarded cryptographic protocol and a very strong initial encryption key.
In short, Point to Point is best to ensure data, but Frame Relay is more cost effective. Analysis of what the requirements are will determine the actual need for one versus the other and defines the ROI (Return on Investment) for the application requirements once defined.