Building MAC Address Tables in a Switch

A switch is a layer 2 device that can be used to forward data from one device to another device. It does that by comparing its database of MAC addresses with the incoming traffic MAC address and port, and where its destination MAC address and port. Although this data can be put in manually (static), it is easier to let the switch learn these MAC addresses automatically (dynamic). This is also useful when the end user devices such as PC connects and disconnects often on different ports.

Basically, a switch first time powered on switch has an empty table. When it received its first frame lets say from PC A, it will save the incoming packet’s MAC address and port on its table. Because it doesn’t know where to send the packets (since the table is empty), it will forward the packet to all ports except from the port it came from. It will then wait for the packet replied by one of the devices attached to it (e.g PC D that is the real destination) and save that MAC address and port. So, next time when another PC sends a packet to PC D it will then go directly to PC D because the switch knows already where to send the packet to based on its table. This goes so on with other packets until all the MAC table and ports are built completely.

In overall, there will be 4 possibilites for forwarding data in a switch. The first one is if there is unicast data from originating device to destination device. If the MAC address of the destination device is known, it will be forwarded to the specific port. Second one, if the data is unicast and the destination device is not known yet in the switch’s database, it will be sent out to every port except from where it came from. The third one is if the data is sent to a MAC address that is known to come from the same port it originates. In this case, this frame will be filtered by the switch and will not be forwarded by the switch to any port. The fourth and last one is multicast or broadcast (remember: broadcast uses ff:ff:ff:ff:ff:ff as destination address). In this case the switch will sent to the designated ports or every port except where it came from.

Internet Cabling

In networking, physical connection is one of the important things to consider. When a problem occurs, the physical connection should be always the first to check. There are several rules in connecting devices that will be discussed in this blog.

The first rule to know is connecting devices of the same type. When we connect devices of the same type we should use a cross over cable. For example, connecting pc-to-pc, switch-to-switch (or to hub because hubs and switch are considered as the same device), and router to routers.

When we connect different type of devices we must use straight trough cables such as router to switch and switch to PC. However, if we want to connect a PC to a router we must use a console cable (or also known as rollover cable) or a cross over cable if use the Ethernet port.

An illustration for this settings is shown in the image below.

Gigabit-Capable Passive Optical Network

Gigabit-capable Passive Optical Network or also known as GPON as one of the optical technologies for telecommunication. This technology allows data communication between the headend and client with a speed of minimum 1 Gbps. According to ITU, GPON can have a coverage of at least 20 km and 60 km with a logical reach. It can also support various bitrate optios such as symmetrical 622 Mbps, symmetrical 1.25 Gbps, and 2.5 Gbps downstream – 1.25 Gbps upstream using the same protocol.

Picture taken from: http://www.fttxtra.com/ftth/gpon/extending-gpons-reach/

One of the reasons why a telco company implements this technology is because of the speed they can get with a long distance. This is the advantage than using LAN which can only have a switch/router with a coverage of 100 meters only. By using GPON, the system needs only to have 1 OLT (typically 4 or 8 ports), a splitter and ONT units. Of course the business model for this system is for corporate or high rise buildings. It is not feasible to put this technology in the residential area because of the high cost (except combined with other technology). Typically one ONT can be installed into a router or switch to cover a smaller area of coverage such as a floor in a building or a cluster of houses, or into another smaller ISP.

As said above, the basic system of GPON consists of OLT, Splitter and ONT. OLT is the headend of the system. ONT is installed on the client and is a passive unit. It works just as a hub. The splitter is used to split one core into many lines and is a passive unit. This is used when multiple ONTs needs to be connected to only a single port. There are currently a 1:2, 1:4, 1:8, 1:16, 1:32, 1:64 and even 1:128 splitter. If a 1 Gbps link (OLT-splitter link) is splitted to 4 customers using a 1:4 splitter, than each link (splitter-ONT link) of the users are 0.25 Gbps (assuming evenly distributed).

I hope this short blog can make you aware of the basic GPON technology.