Connecting Computers

Computers only need a network if they want to connect to other computers, including the Internet and computers in a local area network (LAN). There are various aspects to the connection, including how the signals go from one computer to another, and how computers can find each other once they are part of a working network. We will cover the various aspects in this chapter so that we can connect computers together in a working network.

This chapter covers the applications we’re likely to use, and an overview of the core concepts of TCP/IP networking. We’ll start with the commands as these are succinct.

Commands to connect to computers and Android devices

Connecting to computers

The ubiquitous command (and protocol) to connect to another computer over a TCP/IP connection is ssh. ssh is available for each of the operating systems and platforms we use.

• Linux: ssh is preinstalled.
• OSX (Mac computers): ssh is preinstalled
• Windows: Either use putty or install cygwin and then configure it to include ssh` if it’s not included in your setup.
• Android: ConnectBot is a reasonable Android app that provides basic ssh capabilities. There may be better alternatives.

ssh requires details of the device we want to connect to. We can also provide the username we want to use when we’re connected. Here’s an example: ssh pi@1.1.1.1

We will be asked for the password of that user account. For the current RACHEL images this is rachel. For Raspbian images it’s set to raspberry so try that if you can’t connect using rachel as the password.

Connecting to Android devices

The core command is adb provided in the Android SDK. For Ubuntu this is also available as a Debian package. sudo apt-get install android-tools-adb to install this package.

adb shell connects us to an Android device that has debugging enabled and that’s currently connected using a USB cable. It’s possible to connect using Wi-Fi with some extra configuration.

Commands to transfer files

It’s useful and sometimes essential to be able to transfer files between computers over a network. There are various well-proven commands available. These include scp and rsync on Linux computers (also available for Windows in cygwin) and adb push & adb pull for Android devices connected to a computer.

rsync is particularly useful as it can restart partial downloads. It supports compression while transferring files to reduce the network traffic, which may also reduce costs when the price is based on the bytes transferred.

wget and curl are also worth considering, and I’ve found curl works very well when downloading RACHEL images over the Internet. Here’s an example of using curl to download one of the images:

curl http://rachelfriends.org/rachelpi_64EN/rachelpi2_64EN_3.4.img --retry 999 --retry-max-time 0 -O -C -

Signals and Cables

There are two established ways to connect computers. These are using radio waves and physical cables. Physical networks are almost always using technology called Ethernet. Radio waves vary in frequency, tere are several ‘bands’ of frequencies used for Wi-Fi networks, and others for mobile phone networks such as 3G, CDMA, and 4G (also known as Wi-Max). We seldom need to be concerned with the details unless there are conflicts between several Wi-Fi networks.

Ethernet uses physical cables which connect devices to either a network switch or a network hub. We will cover these in the next section.

Wi-fi networks use one of the 802.11 standards. Again, we seldom need to worry about the details. We mention these standards in case you ever need to troubleshoot connection problems.

Wi-Fis, switches and hubs

One thig is common to all networks, they need some at least one physical device to make the connection to other devices, even when the connection is over Wi-Fi or a mobile network.

• Wi-Fi: There are two variations of Wi-Fi network. One connects between two computers and may be called an ad-hoc or point-to-point network. This can be useful to share information and materials between those two devices, however it’s a very restricted network as other computers can’t participate so we seldom use them. More common and often more relevant for projects such as RACHEL is an Infrastructure network where one device is the master and controls the network. The RACHEL server is generally configured to be the device that controls and provides a Wi-Fi network.

In some locations there may already be a device configured to provide the Wi-Fi network, if so, then we recommend you configure the RACHEL device to be part of that network rather than trying to provide another network. We also recommend you allocate a fixed IP address for the RACHEL server so that other computers, and the users, can find that server easily and consistently. Allocate an IP address that isn’t part of the pool of loan addresses (provided by the DHCP service). This will reduce the chances of conflicts within the local network.

Finding each other

Each computer and device on a network will need a known IP address in order to communicate with each other and with computers beyond the local network. Sometimes we may end up using an IP address directly, for instance in the web browser of an Android tablet to connect to the RACHEL server.

IP addresses

There are two main standards for IP network addresses, the older ubiquitous IPv4 addresses, such as 192.168.1.102 and the newer IPv6 addresses. We will only cover the IPv4 addresses here as they’re the primary format used by the various computers and other devices we use.

This book aims to explain the essential aspects simply. This means we decided to exclude some precsion and some accuracy of the entire topic. There are plenty of materials available elsewhere that should provide more accurate extra information. Cisco’s CCNA is a well respected qualification for people who want to understand networks well.

IP addresses need to be unique for each device on a network. Otherwise there may be problems that stop one or more of the computers from communcating as they will be competing for what is intended to be a unque address. Computers aren’t very smart about trying to resolve the problems, an administrator of the computer network may need to make changes to the network configuration and/or to the network settins of devices trying to use that network.

IP addresses can be provided on loan, preallocated, or configured on the local computer or device. Typically IP address are configured on servers, such as the RACHEL server, and on network devices such as the Wi-Fi router.

Most other devices request an IP address when they start using a network. They also need to ask if they want the loan to be extended, as each loan has a finite period. Loans vary from several minutes for very crowded networks such as an Internet Cafe, to many months where the administrator doesn’t expect much to change in terms of the computers using that network.

If our assumptions when configuring the loan period are incorrect some devices may not be able to connect to the network.

The addresses are loaned by a service called DHCP. In some cases, such as when the RACHEL server provides the Wi-Fi network, these addresses need to be configured on that server (the preconfigured RACHEL servers are already configured to do so). Wi-Fi routers generally have a web-based administrator’s interface where the set of addresses are configured.

Common IP addresses for RACHEL Servers

There are at least 3 IP addresses which the project has used in RACHEL Pi builds. These include:

• 1.1.1.1
• 10.10.10.10
• 192.168.1.10

If you’re computer or device is connected to the Wi-Fi provided by RACHEL Pi you can use your IP address to determine (or guess) the IP address of the RACHEL Server. You may be able to check your IP address in the GUI, or on Windows, OSX and Linux, from the command line using either ipconfig /all (on Windows) or ifconfig on other operating systems.

Local and foreign networks

Conceptually there are local addresses and foreign ones, somewhere else on another network. Generally we want our computers to share a common, single logical local network. This means that most of the IP address will be the same, this is always the first part of the address, reading from the left. For instance they may all begin with 192.168.1 where only the final part is unique.

Getting a RACHEL Server running on an existing network

If there is an existing computer network available we may want to configure our devices to share that network. To do so we would need to know various parameters such as the name of the Wi-Fi network, any password, etc. For devices where we want to assign a fixed IP address we will also need to know what IP address to use, and other details such as the subnet, the default gateway and the DNS server. We will cover each of these later in this chapter.

Raspberry Pi Model A’s

Raspberry Pi model A’s don’t have an Ethernet port, so they’ll need to connect to the network using a USB Network adapter. Typically this is a Wi-Fi adapter such as the WiPi. The current RACHEL images configure the WiFi adapter as a network HotSpot, where it assigns IP addresses. Typically the existing network will also have a machine (typically the router to the Internet) that assigns IP addresses. Generally you’ll want to reconfigure the RACHEL so that it’s no longer a HotSpot, instead it would use an IP address from the existing Network.

Raspberry Pi Model B’s

Raspberry Pi Model B’s have an Ethernet port. Sometimes we may add an additional USB Wi-Fi adapter. Again the current RACHEL images configure the WiFi adapter as a network HotSpot. This means the Pi (and therefore RACHEL) can potentially be used on either network provided we also configure the Ethernet port so it has an IP address of the same network as the existing (non RACHEL) network.

Obtaining an IP address

The network port could use a static IP address - one that you configure the Raspberry Pi to use, or it could be configured to request and receive an IP address chosen by a service called DHCP. This service will be running on the machine that assigns IP addresses.

Finding a Raspberry Pi on the local network

If you assigned a static IP address to a Raspberry Pi, you know enough to reach it without further information. for instance you could SSH in to the operating system running on the Pi (I’m assuming SSH has been enabled and configred, as it is on Rasberian and RACHEL images), and you could connect to RACHEL with a web browser using by adding the prefix http:// to the IP address of the Pi.

If you don’t know the IP address of the Pi, there are various ways to find it. These include:

• Using the Web interface of the router, which will often list which DHCP IP addresses are being used.
• A network scan using a command-line utility (nmap, or ping), etc.
• Connect a keyboard and monitor to the Pi, login (probably using the pi account) and run ifconfig from the command line, then search for the IP address in the results. (there are ways to refine this command e.g. piping the output as follows ifconfig | grep inet and/or specifying the interface e.g. eth0
• Try using ascending IP addresses from the range that’s been assigned, while this can be very time-consuming and tedious, it generally works eventually and can enable you to find the RACHEL software when no other tools or techniques are available

Each network adapter has a low-level address, called a MAC address. The first half identifies the manufacturer of the network adapter. With practice we can learn to recognise the prefix for network adapters used by the Pi. TBC add commonly used prefixes for the Ethernet and WiPi adapters.

Subnets

The subnet is a filter that matches the common part of the addresses, at least for the purposes of this book and our projects. The popular term is the subnet mask, which can be represented in several forms. These include the .dot notation e.g. 192.168.1.x and /24 which represents the number of bits in the network address that remain constant for each address in this network.

Default Gateways

Computers address each other directly when they share a network. To reach computers in other networks they forward all their requests to a network address known as the ‘default gateway’. This may be a device such as a router, or another computer such as a Raspberry Pi. That device is responsible for trying to find a route to connect the computer making the request to the relevant remote device, often another computer such as a web server.

By convention, default gateways often have an IP address that either ends in .1 or .254 Computers need to be told what the address of the defult gateway is. They get this information with their IP address.

Names, URLs and IP addresses

Humans generally prefer to use names than numbers to address things. However, computers including their web browsers actually use the IP address as a number. They need to obtain the correct IP address for whatever name is being used. There are several ways to do so, the most common are either using a DNS Server or a small text file locally on that device. When we use a mobile network or an existing Wi-Fi network they often ‘know’ where to find the DNS Server which then performs the mapping between the name and the IP address that matches that name.

Sometimes we may want to configure some of our computers to override several of the names being used so that informaton can be provided within the local network rather than needing an Internet connection. These local settings can help provide updates of software and materials. They can also help reduce undesireable use of potentially expensive and scarce Internet connections, for instance to restrict access to sites such as social networks (Facebook, Twitter, Instagram, etc).

Configuring a RACHEL Server

• TBC: until we add the relevant material to this book, a good place to obtain the information is on the wiki of the ka-lite project https://github.com/learningequality/ka-lite

Advanced topics

Sometimes we need to do some more complicated networking in order to get our work done. One example is connecting a Raspberry Pi (running as a Wi-Fi hotspot) to the Internet through a laptop connected over Wi-Fi. I have had to use this approach to update RACHEL Pi, for instance to update the kernel files so the image will run on a new model of Raspberry Pi.

Reverse-proxy on a computer

When a computer is configured as a reverse-proxy, it allows connected devices to use it to connect to another network. The computer also needs to be configured as a DHCP server (to provide the Pi with an IP address to its Ethernet port, the Wi-Fi adapter on the Pi is being used as a Wi-Fi hotspot).

The configuration of a reverse-proxy is specific to each desktop operating system. TBC add some links here.

Here’s an overview of how to configure a reverse-proxy on OSX http://makezine.com/2012/07/16/use-your-mac-laptop-as-a-wireless-proxy-for-raspberry-pi/ ##Further reading I found a very helpful guide on IP protocols and how all this stuff works, at http://ipprimer.com/#/ perhaps you’ll find it useful too?