In the early days of Networking, a computer network was meant to carry only data traffic. As time went on, VOIP, or Voice-Over-IP technologies were invented, computer networks had to be designed to also carry Voice traffic. When a single network fabric is carrying both Data traffic and Voice traffic, it is referred to as a Converged network.
Voice traffic and Data traffic
When designing a network, it is important to determine what type of traffic will be traversing the network. Based upon the type of traffic, the network can be built to emphasize up to four different network characteristics:
- Bandwidth – The max capacity of a particular network. Measured in “bits per second” (bps, kbps, mbps, gbps, etc).
- Throughput – The current rate of traffic being pushed through a network. Measured in “bits per second” (bps, kbps, mbps, gbps, etc).
- Latency – The time it takes to get from one point in your network to another. Measured in “milliseconds” (sometimes measured as the Round Trip Time, or RTT).
- Jitter – The variance in Latency in your network between your “busy times” and “slow times”.
These four attributes can be prioritized in order to optimize your network for the type of traffic it is meant to carry.
A network carrying mostly Data traffic will predominately care about Bandwidth, and only moderately care about Latency and Jitter. If you are transferring 100 gb (100,000,000,000 bits) of data at a rate of 10 mbps (10,000,000 bits per second), it will take you 10,000 seconds (2 hours, 45 minutes) – an additional 1-3 seconds due to latency and/or jitter will largely be unnoticed.
A network carrying mostly Voice traffic will predominately care about Latency and Jitter, and only moderately care about Bandwidth. If you are speaking to someone in “real time”, the size of voice packets is not particularly large (i.e., doesn’t require high throughput), but having each “word” get to the other end as fast as possible is crucial. Imagine how frustrating a phone call would be if each word were delayed by a few seconds. The effect would be very frustrating.
To accommodate the different priorities, the legacy solution was to build two independent networks: one optimized to carry Data traffic, and the other optimized to carry Voice traffic:
This allowed for separation of Voice and Data traffic, but required twice the gear to build out the network.
As network gear progressed in performance and functionality, however, the industry was able to run both Data and Voice traffic on the same network devices (routers, switches, etc). This was referred to as a Converged Network – a single network fabric that can carry both Voice traffic and Data traffic:
Of course, the network fabric still has to prioritize different characteristics for Voice or Data traffic, therefore the network must be able to distinguish the Voice traffic from the Data traffic.
The primary method for the network devices to identify different types of traffic is to use different IP Networks. The VOIP phones (Voice Over IP) would be assigned a particular set of IP addresses, and the PC’s would be assigned another set of IP addresses.
Then, the network gear would apply different priorities to traffic based upon the IP addresses it came from using a feature known as QoS, or Quality of Service.
Converged Network using VLANs
The image above uses two different switches and two interfaces on the router for the Data and Voice traffic in order to visually represent the different IP networks. But the same effect can also be accomplished with one switch and one router interface using VLANs:
In the image above of a Converged Network, both VOIP phones and PC’s are plugged into the same switch, but are still logically separated in different IP networks using VLANs: a Voice VLAN and a Data VLAN.