Monday, October 26, 2015

Week 8: Day 024 - Installing a Physical Network #6


Hello, it's a new week, and there's not that much left in this chapter, thankfully! I'm really eager to get onto the TCP/IP part of the textbook (which is next.) Anyways, today we're going to go into some NIC related stuff, fun!

So as we know from before NIC stands for "Network Interface Controller" and the textbook characterizes this as the "final part of any physical network". All UTP Ethernet NICs have the RJ-45 connector, which is connected from the NIC to the switch. Now, when it comes to fiber-optic NICs, they come in variety. Manufacturers like to use the same connectors for the multiple standards, which causes confusion. So a 100BaseFX card used for multimode with an SC, and an identicle card for single-mode also uses an SC connector. The only way of telling them apart is the documentation!

When you're looking to buy a NIC, Mike Meyers suggests buying from brand names such as Intel, because they're reliable, have more features, and are easy to return if they're defective. He also points out that, replacing missing drivers on a brand name NIC is also easier (which I would imagine.) But the NIC you choose depends on the network. If you can afford it, get multispped cards, and stick withthe same model. The more models you're using, the more driver discs you'll be carrying around (which sucks.) So all in all, I would have to agree with this assessment, go with the brand name NICs!

The physical connections in a computer are not magical, so if the NIC is not connected to the computer, the NIC will not work! Fortunately, most of the time the NIC is built in to the PC you purchase. Usually there will be a couple expansions slots in the computer. One of them is older, but more common, and it's known as PCI (Peripheral component Interconnect.) The newer one is PCIe (PCI Express) which are used more widely nowadays. They come in one-lane or two-lane varieties, and I have no clue what that is haha. If you're so lazy that you won't open up the PC case, you can be lazy, and get a NIC with USB and PC Card ocnnections. However USB 2.0 is not as good as gigabit ethernet, so it's a dumb idea to use that, and PC Card is exclusively for laptops. But a USB 3.0 NIC are brilliant and handles speeds up to 5 Gbps, so there's no problem with those. These USB NICs come in handy when you must troubleshoot whether the machine you're dealing with has a bad NIC or not.

Next, there are the NIC drivers that must be installed onto the OS of the machine. It's very easy, just put in the CD, go through the installation, and you're done. Sometimes, it might even already have the driver installed. Every OS has its own way of identifying the existence of the NIC. Then once you've done that, an important thing to note is that two NICs can actually be used on the same computer! The process to achieve this is called "bonding" or "link aggregation". Basically it doubles the speed between a machine and the switch, so happy days.

Finally, there's link lights. Every UTP NICs have LEDs (Light-emitting diodes) which give information about the status of the link between the switch and the machine. Even though you may know that it's an LED, call it link lights. NICs have four link lights and they can be any color. They literally tell you what may be wrong with the system's connection. More specifically, a link light tells you that the NIC is connected to a switch, and the switches themselves also have link lights. If one of the computers on the network are telling you that it's disconnected, consult the link lights first. When it's functioning correctly, the link light is on and is steady. This means NO flickering, just on. Then there's the "activity light", which turns on when the card detects traffic on the network. The light is great for detecting problems, because the connection light is sometimes fraudulent in it's claims. If the connection light says it's fine, try to copy files, create network traffic somehow, and if the activity light doesn't flicker, then you know there's a problem. On older NICs there is the "collision light", which flickers when there are collisions on the network. It may not be important, but it could be on the exam. There is no standards on the NIC lights, so manufacturers created many varieties of it. When you see a new NIC with different lights, figure out what each one means. It's always going to be the same: link, activity, and speed. When it comes to fiber-optic NICs, they don't have lights! This make troubleshooting a harder task. To add on to this, the cabling itself is very delicate. People that work with fiber have testers of their own, as I said earlier in the chapter. Honestly, it probably doesn't matter much to me, but it's something to note. That ends the section on NICs, tomorrow will be a shorter entry, since I will be doing the final section! On that bombshell, thanks for reading, and goodbye!


1 comment:

  1. You got me curious with your mention of single lane and double lane, so I did what I usually do in these situations, turned to Wikipedia:

    https://en.wikipedia.org/wiki/PCI_Express

    I had no idea that the myriad of connectors I've seen on motherboards over the last several years were all the same standard (all PCI interfaces looked the same).

    This article makes a good read, I highly recommend it.

    ReplyDelete