Routers are networking devices that direct traffic between subnetworks on the internet. The word "router" is pronounced like "doubter."
The computer you're using is almost certainly part of a wired (Ethernet) or wireless (WiFi) local network connecting the computers in your house or your school. The networking hardware in your computer can only talk to other computers on the same local network. How it is that you can use the Internet to connect to any computer in the world?
One (or maybe more) of the computers on your network is a router: a special purpose computer that passes messages from one network to another. If you have a router at home, it probably connects your local network with a larger network belonging to your Internet Service Provider (ISP), generally your telephone company or your cable TV company. The job of a router is to pass messages from one network to another until it reaches the local network of the computer you're trying to reach. Your home router connects only two networks, your local net and your ISP's net. Your ISP has many routers, each of which connects several networks.
The Internet isn't a network of computers; it's a network of networks.
How do the routers know where to find the computer you want? Every computer on the Internet (including smart phones, routers, Internet-enabled appliances, etc.) has an address (or more than one, if it's a router), like a postal or email address, called its IP address. IP stands for Internet Protocol. (A protocol is a set of rules for communicating on the net; IP is the protocol that specifies how a router handles a request for another IP address.) You can find out your computer's IP address by asking a search engine:
It's a search engine, like Google or Bing, except that it doesn't track your searches to sell you to advertisers as the others do. You saw it in Unit 3 Lab 5.
What's with the "?q=" in the URL https://duckduckgo.com/?q=my+ip+address?
The "?q=" tells the duckduckgo.com server to run the query "my+ip+address," which returns the client's (your) IP address. (The "q" is for "query" as we are querying the server for information.)
What's an undecillion? One undecillion is 1036, a billion billion billion billion (a 1 with 36 zeros after it).
Watch this video from code.org:
IP Address Hierarchy
The following was constructed from Dan's internet PPTX and Brian's to original IP Page --MF
Next year, the text and graphic below should be changed to snap.berkeley.edu at 220.127.116.11. See Dan's PPTX files on Internet (one and two) to modify. --MF
I recognize that it's not as simple as "just changing to snap.berkeley.edu" because servers have all these funny rerouting things going on (and york might actually be where snap is, etc.). Brian, we should talk about this. (Or just leave me a comment here or in my inbox about it. Thanks!) --MF
snap.b.e is complicated; it's translated to byob.eecs.b.e which is in turn translated to eecs.b.e/~bh/byob. It'd be a little easier to talk about bjc.b.e, which is called bjc throughout the translation steps. But in the end they both are the same computer as eecs.b.e, namely 18.104.22.168. So you could legitimately ignore all that referral stuff and just say snap.b.e is 22.214.171.124. If you want a nearby computer for comparison, the one on my desk, abbenay.cs.b.e, is 126.96.36.199. Where did 188.8.131.52 come from?--bh
Is "Recall" here true? Have we talked about hierarchies before? They haven't taken biology yet, maybe.--bh
Yes, on page 1 of lab 1. --MF
Recall, a hierarchy is an ranking arrangement (like species taxonomy) with the broadest or most powerful rank at the top. It's often depicted as a pyramid.
Like domain names, IP addresses are hierarchical.
For example, consider the IP address for york.cs.berkeley.edu, which is 184.108.40.206. In the picture below, the subnetwork for the 128.32 domain includes two subnet numbers: 189 and 156. Subnet 156 has two hosts: 218 and 226.
This york.cs.berkeley.edu hierarchy image is a simplified model.
Each of the four numbers in a typical IP address today is an eight-bit byte with a value between 0 and 255:
A 32-bit IPv4 (the "v" stands for "version") address is big enough to support 232 computers, which is about four billion (4 · 106). There are more than seven billion people on Earth, so there aren't enough IP addresses to go around.
Why 232? There are 32 bits, and each bit can be one of two possible values (0 or 1). So, there are two possibilities for the first bit, four (that's 22) possibilities for the first two bits, eight (that's 23) possibilities for the first three bits, 16 (that's 24) for the first four bits, and so on up to 232 possibilities for thirty-two bits.
The long-term solution is to increase the length of an IP address. The new IP addresses are 128 bits wide, which is enough to support 2128 computers:
2128 is about 1038. For contrast, there are about 1050 atoms in the Earth and an estimated 1029 stars in the observable universe. So even if the Internet is extended to include other planets or space aliens, we'll still have enough addresses with IPv6 (but not if every atom had its own IP address).
Different protocols can be used at the same time; both IPv4 and IPv6 are in use today. This is a great example of the importance of abstraction. Most traffic (even traffic across new routers capable of IPv6) still uses IPv4. This is because in order to make an IPv6 connection, all the routers along the way must recognize IPv6. Old routers are replaced all the time, and eventually every router on the Internet will use IPv6.
Your IP Address
Look up your current IP address using a search engine (as described at the top of the page). Is your address in IPv4 or IPv6?
Load your U4L1-HttpBlock project.
Write a Snap! reporter in the Sensing palette that reports your IP address. You can't, unfortunately, do it by asking DuckDuckGo as you did earlier, because you'd have to use the proxied http:// block, and so you'd get back the address of the proxy site instead of your own. So instead, use alonzo.herokuapp.com/ip-address, which allows connections from the unproxied http:// block. You'll use this block in Lab 5, to write a program that checks the weather where you are, without having to ask you where you are; it can tell (pretty closely) from your IP address.
Most likely, the router in your home or school uses a protocol that allows all the computers on a local network (for example, in one building) to share one IP address on the Internet. The router that creates the local network gives each computer a local address. For example, although the outside world may think someone's computer has IP address 220.127.116.11, that computer itself probably thinks its address is something like 192.168.1.11.
Look up your current local IP address in your system preferences or settings. It's usually under network or internet settings and may be listed with the computer device supporting that connection (wifi, ethernet, wifi, bluetooth, etc.).
The 192.168 domain (the block of IP addresses that all start with 192.168) is reserved for local networks; no computer on the Internet has an address in that range. Another such domain is 10.0.
I've gleaned bits from this as marked. Otherwise it's not going to be included. -MF
OK, except see my comment about end-to-end below. --bh
It's on that page, but it's written generally. Can you take a look (it's in an orange box on TCP) and make whatever changes you think are necessary? --MF
Network people have their own vocabulary. On the AP exam you may see
The end-to-end principle is abstraction, in the particular case of the Internet hierarchy. It says that routers don't have to know anything about the meaning or purpose of the messages you send over the net; all they care about is the IP address at the beginning of the message. Your computer and the server to which you sent the message are the only ones that have to think about what it means.
As a result, when someone invents a new application for the Internet, all the routers in the world don't have to be replaced or reprogrammed — to them it's all just bits.
Robustness is the technical name for the redundancy of paths through the Internet, It means that if part of the net fails, the rest of the net remains connected even if the failed part is in the usual path from one place to another.
Robustness and the end-to-end principle improve Internet scalability, which is the ability of the net to keep working as the size of the network and the amount of traffic over the network increase. (The page Internet 2012 in numbers has some astonishing numbers about Internet traffic from a few years ago.)
packet is a small message. Typically, if you want to send a large message, your computer divides it into small chunks that it sends individually. This is important because if a router that's handling a chunk of your message fails, only that chunk has to be resent, rather than the entire message. The Internet is a packet switching network. (By contrast, land line telephone systems establish a long-term connection from one telephone to another, for as long as the phone call continues.)
store-and-forward network, because each router along the way stores your packet long enough to forward it to the next router.
Another name for a packet switching network is a
lot of protocols! Here's a picture from Wikipedia:
You don't have to know about all of these, but you should know the official names for the four levels of abstraction (layers) in the Internet hierarchy.
There are a
open standards, which means that anyone can look up their details and build them into new hardware and software without getting permission. (For example, here is the IPv4 specification, but don't read it now.) Before the Internet, there were several different network protocols that were secrets belonging to particular manufacturers. So if you had a particular brand of computer or router, it could talk only to other computers of the same brand. Open protocols are another reason why the Internet is scalable.