5.1.1 Local Area Network and Wide Area Network

A computer network can be physically classified as either a local area network (LAN) or a wide area network (WAN). A LAN is a group of interconnected computers within a building or a campus. A WAN is a collection of LANs that are spread out over a large area. A WAN within a city is also called a metropolitan area network (MAN). A LAN typically belongs to one organization and has high-speed connections, such as Fast Ethernet (100 Mbps) or Gigabit Ethernet (1000 Mbps). A MAN or WAN typically belongs to different organizations and has relatively low-speed connections. The Internet is an example of a worldwide public WAN.

5.1.2 The Cisco Three-Tier Enterprise Network Architecture

Large organizations with huge computer networks typically adopt Cisco's three-tiered enterprise network architecture, shown in Figure 5.3. From top to bottom, the tiers are Core, Distribution, and Access.

• The Access layer provides the network access to end-user devices, such as computers, printers, or servers. The typical networking devices are switches, which connect all the devices in either a star network topology or a tree network topology (which consists of several star network topologies connected in a hierarchical structure).
• The Distribution layer bridges the gap between the Access layer and the Core layer. The typical networking devices are routers, which connect all the switches in the Access layer to the Core layer. The Distribution layer also provides the network connection with redundancy by using mesh network topology.
• The Core layer connects all the distribution routers together and provides the Internet access. The core layer is also known as the backbone. The typical networking device is a gateway router, also called the default gateway. Large organizations may have more than one gateway router to provide the Internet connection with more bandwidth and redundancy.

5.1.3 Key Network Components

Figure 5.3 shows the following key network components: switches, routers, a default gateway, and various servers such as DNS and DHCP servers. For a Windows computer network, you will also need a Windows Internet Name Service (WINS) server.

• A switch is a device used to connect users’ client computers and usually has a large number of the same type of interface, called ports. A switch can have 8 ports, 16 ports, 64 ports, or even 128 ports.
• A router is a device to connect different networks, not users’ computers, and usually has a few interfaces of different types, such as Ethernet, Fast Ethernet, Gigabit Ethernet, serial, Fiber Distributed Data Interface (FDDI), Asynchronous Transfer Mode (ATM), and Integrated Services Digital Network (ISDN). Routers are so called because they also perform routing functions to find the best route to a destination.
• The gateway, or default gateway, is where the local computer network connects to the Internet. The gateway is also the place where the network proxy server, network policy, and network firewall are commonly implemented.
• The DNS is another key component. The computers work based on IP addresses, but we humans are not good at remembering numbers, so we prefer names. DNS servers translate the human-understandable domain names into machine-understandable IP addresses, and vice versa.
• The DHCP server automatically provides the network configuration information for your computer. The following are four minimum pieces of information you need to configure your computer to access the Internet:
  • IP address: To identify the device
  • Subnet mask: To identify the subnet of the device
  • DNS server: To know where the DNS server is
  • Default gateway: To know where the gateway router is

5.1.4 Traditional Networks vs. Software-Defined Networking

Traditionally, computer networks are composed of switches and routers. Each switch or router has two planes, a control plane, and a data plane (or forwarding plane), as shown on the left in Figure 5.4. The control plane determines where to send traffic data, and the data plane executes these decisions and forwards traffic data. In this scenario, you will need to configure each switch or router individually. This approach is therefore time-consuming and error-prone, which makes it difficult to cope with the scalability problems of growing computer networks.

Software-defined networking (SDN) is an emerging technology that has become a new key buzzword in the computer networking/IT industry. With SDN, the control planes of switches and routers are taken away from individual devices and managed centrally using a dedicated server, as shown on the right in Figure 5.4. SDN is a software layer that sits on the top of existing computer networks. The server controls the traffic centrally by programming the switches and routers directly. This brings several benefits.

• Central management: You can configure, monitor, and troubleshoot the network all from the controller. You can also view the complete network topology from the controller.
• Lighter-weight network devices: Because all the controls are performed centrally, you can use slimmed-down switches and routers to reduce network costs.
• Network virtualization: SDN helps to virtualize the networks. Virtualization of networks has brought a revolution in the IT industry to virtualize storage and computing entities, which have played a key role in efficiently utilizing resources.

The most commonly used SDN communication protocol is OpenFlow, which is an open standard managed by the Open Networking Foundation. (You'll learn how to obtain and get started with it later in the chapter.) OpenFlow needs to be implemented in both switches/routers and the controller. OpenFlow carries messages between switches/routers and the controller. To control network switches/routers, the controller will use OpenFlow to push rules, such as forwarding rules and security rules, into switches/routers so that they can make decisions when network traffic hits them. Switches/routers need to maintain such rules in an OpenFlow table. Rules are also called flows and are stored in flow tables. To monitor switches/routers, the controller uses OpenFlow's various request and response messages to fetch the statistics information and events messages to update the controller about changes or failures that occur at switches/routers.

Floodlight is a popular SDN controller project. (You'll also learn how to obtain and get started with it later in the chapter.) It is a Java-based, Apache-licensed OpenFlow controller that you can use to help build a software-defined network. Floodlight is developed by an open community of developers, and it is simple to build and run. Floodlight supports a broad range of virtual and physical OpenFlow switches. Floodlight can handle mixed OpenFlow and non-OpenFlow networks. Floodlight is designed to be high-performance; it is multithreaded from the ground up, and it supports the OpenStack (link) cloud orchestration platform.

OpenDaylight is another Java-based open source SDN controller project. (You'll also learn how to obtain and get started with it later in the chapter.) It is a baseline project upon which many other controllers are built. The OpenDaylight project is conducted by the Linux Foundation. The goal of the project is to promote software-defined networking and network functions virtualization (NFV). NFV is another hot topic at the moment. SDN is closely related to NFV but is different. SDN focuses on separating the network's control plane from the forwarding plane and provides a centralized control of the network. NFV focuses on improving and virtualizing the network services themselves, such as DNS, caching, load balancing, firewalls, intrusion detection, and so on. Both SDN and NFV aim to advance a software-based approach to networking to make the network more scalable, more agile, and better supported.

In summary, SDN provides traffic programmability, agility, and the ability to create policy-driven network supervision and to implement network automation. It also allows the creation of a framework to support more data-intensive applications such as big data and virtualization. SDN can reduce both the capital expenses (CAPEX) of network equipment and the operational and maintenance expenses (OPEX) of a network. As a result, more and more companies are starting to adopt SDN.

For more details about SDN, please visit the following resources:

https://en.wikipedia.org/wiki/Software-defined_networking

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

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

https://www.cisco.com/c/en_uk/solutions/software-defined-networking/overview.html

https://www.juniper.net/uk/en/products-services/sdn/

https://github.com/mininet/openflow-tutorial/wiki

http://www.projectfloodlight.org/floodlight/

https://www.opendaylight.org/

http://sdnhub.org/tutorials/

5.3.1 Java UDP Client-Server Programming

The User Datagram Protocol is a Transport layer protocol. UDP provides a connectionless, unreliable, but simple and fast service. Therefore, UDP is mostly used in time-critical applications such as Internet telephony, web broadcasts, or online video, where sending data without delay is more important than sending data without error.

Example 5.6 is a simple UDP client-server program. Example 5.6A, UDPServer1.java, is the server program, and Example 5.6B, UDPClient1.java, is the client program. In this example, the server program needs to run first, waiting for the client's connection, and the client program needs to run second so it can connect to server. In the client program, users need to type in a one-line sentence that will be sent to the server. When the server receives this data, it will echo it back to the client. Finally, the client will print out the server's response.

The following is the UDPServer1.java program's execution logic:

begin
    create a Server Datagram Socket on port 3301
    create a Datagram packet for receiving data
    receive data from a client
    print client's IP address and port number
    create another Datagram packet containing client's data
    send packet to client 
    close Socket
end

The following is the UDPClient1.java program's execution logic:

begin
    create a BufferReader associating with keyboard
    create a client Datagram Socket
    read in a line from keyboard
    create a Datagram packet containing that line
    send packet to server
    create another Datagram packet for receiving data from server
    receive data from server 
    print server's response
    close Socket
end

Figure 5.10 shows the compilation and execution of the UDPServer1.java program and the UDPClient1.java program. In this example, the server runs on localhost (127.0.0.1), port 3301; the client runs on localhost (127.0.0.1), port 3024. Please note that you will need two separate console windows to run the UDPServer1.java program and the UDPClient1.java program.

5.3.2 Java TCP Client-Server Programming

Transmission Control Protocol is also a Transport layer protocol. Unlike UDP, TCP provides connection-oriented, reliable, streaming services. Because of TCP's error detection, error correction, and traffic flow control, TCP has been widely used in web, email, FTP download, and Telnet services, where the integrity of the data has the highest priority.

Example 5.7 is a simple TCP echo program. TCPServer1.java is the server program that provides an echo service. TCPClient1.java is the client program that sends a line of data to the server and prints the server's echo reply on screen.

The following is the TCPServer1.java (Example 5.7A) program's execution logic:

begin
    create a server TCP Socket on port 3301
    create a client Socket waiting for client's connection
    print client's IP address and port number
    create a BufferReader for receiving data from client
    create a DataOutputStream for sending data to client
    send data to client 
    close client Socket
    close server Socket
end

The following is the TCPClient1.java (Example 5.7B) program's execution logic:

begin
    create a BufferReader associating with keyboard
    create a client TCP Socket associating with a TCP server
    print server's IP address and port number
    create a DataOutputStream for sending data to server
    create a BufferReader for receiving data from server
    read in a line from keyboard and send it to server
    create a Datagram packet containing that line
    receive data from server and print server's response
    close Socket
end

Figure 5.11 shows the compilation and execution of the TCPServer1.java and TCPClient1.java programs. Again, you need to run TCPServer1.java first in a console window and then run TCPClient1.java in another console window. In this example, the TCP server runs on localhost (127.0.0.1) and on port 3301, and the TCP client runs on localhost (127.0.0.1) and on port 3020.

5.3.3 Java Multithreaded EchoServer Programming

The echo server in the previous section runs on a single thread and therefore can accept only one client at a time. Example 5.8A shows a multithreaded echo server, which allows the echo server to accept multiple clients simultaneously and echo back the message from each client separately.

Example 5.8B shows the corresponding echo client.

For more information on Java socket programming, see the following resources:

https://docs.oracle.com/javase/tutorial/networking/sockets/index.html

https://www.javaworld.com/article/2077322/core-java/core-java-sockets-programming-in-java-a-tutorial.html

https://o7planning.org/en/10393/java-socket-programming-tutorial

5.4.1 A Java HTTP/HTTPS Client

Example 5.9 shows a series of Java HTTP clients—web clients. First, Example 5.9A is a simple client using the URL and URLConnection classes. In this example, it opens the web page specified by its uniform resource locator (URL), in this case www.google.com/. Figure 5.12 shows the compilation and execution of the program.

Example 5.9B is a variation of the previous Java program, which uses HttpsURLConnection() to connect to a secure web site, using the HTTPS protocol.

Example 5.9C is a version of the same program, but it also prints the digital certificate information from the HTTPS server.

HTTP/HTTPS clients can use either the GET or POST method to talk to HTTP/HTTPS servers. The default is the GET method, which is used to retrieve information from the server. The POST method is for sending information to the server, such as sending login details, submitting forms, searching keywords, and similar tasks. Example 5.10 shows a simple Java HTTP/HTTPS client using the GET method to get information from the server, similar to the previous example. This version of the program opens the web page specified by its URL, in this case https://docs.oracle.com/javase/tutorial/. Figure 5.13 shows the compilation and execution of the program.

Example 5.11 shows a simple Java HTTP client using the POST method to send search results to the server. In this example, it opens the web page https://www.amazon.co.uk/s/ref=nb_sb_noss_2? and searches for the keyword java. Figure 5.14 shows the compilation and execution of the program. Because of the output length, only the first 1,000 characters are printed. You can also search in another web site, such as Google; just uncomment the following two lines in the code:

         //String website="https://www.google.co.uk/search?";
         //String urlParameters="q=java"; 

Example 5.12 shows another Java HTTP example using URL class. It parses the URL of a web page and shows the corresponding information such as the protocol, host, port, path, and query. Figure 5.15 shows the compilation and execution of the program.

5.4.2 A Java HTTP Server

Example 5.13 shows a simple Java HTTP server program, using ServerSocket classes. It runs on port 8088, and when a web browser client is connected, it displays Hello World! in the browser. Figure 5.16 shows the compilation and execution of the program. To view the web page, open a web browser and type the IP address of your computer and port 8088, as shown in the figure.

5.4.3 Java Multithreaded HTTP Server

Example 5.14A is a multithreaded version of the previous HTTP server. In this version, it first creates a ServerSocket object on port 8088 and then uses a while loop to listen to client requests. Each time it accepts a client connection, it generates a separate thread dealing with that connection. Inside the thread, within the run() method, it reads all the lines of the HTTP request message from the client and sends back a Hello Multithreaded HTTP Server! message.

Example 5.14B is another multithreaded version of the earlier HTTP server. In this version, it finds out what file does the client request, finds and opens the file, reads line by line until finished, and also sends each line to the client accordingly. To test the HTTP server, just open a web browser and type the IP address of your computer and port 8088, exactly the same as in the previous figure.

5.7.1 Getting Started with OpenFlow

The simplest way to get started with OpenFlow is to use Mininet. Figure 5.18 shows an interesting GitHub site (https://github.com/mininet/openflow-tutorial/wiki) on the OpenFlow tutorial. It shows how to download and set up the relevant software and shows examples such as Learn Development Tools, Create A Learning Switch, Control A Slice Of A Real Network, Router Exercise, Advanced Topology, and Create Firewall.

To get started, you first need to download the following software:

• Mininet VM Images

https://github.com/mininet/mininet/wiki/Mininet-VM-Images

• VirtualBox

https://www.virtualbox.org/wiki/Downloads

• Xming

https://sourceforge.net/projects/xming/files/Xming/6.9.0.31/Xming-6-9-0-31-setup.exe/download

• Putty.exe

http://the.earth.li/~sgtatham/putty/latest/x86/putty.exe

Then follow the instructions at the following sites to import the Mininet virtual machine file into the VirtualBox software:

https://github.com/mininet/openflow-tutorial/wiki/Set-up-Virtual-Machine

http://mininet.org/vm-setup-notes/

Figure 5.19 shows the Mininet virtual machine in VirtualBox. Start the Mininet virtual machine, and use mininet for the username and mininet for the password to log in. The Mininet virtual machine is based on the Ubuntu Linux operating system.

After login, at the Linux shell prompt ($), you can type the following command to start Mininet and enter its command-line interface (CLI) mode. The sudo command means you run the Mininet as a superuser, or administrator.

$ sudo mn

This will start a simulated network with minimal topology, which includes a controller (c0), a switch (s1), and two hosts (h1 and h2), as shown in Figure 5.20.

You can also start with other topologies by using the --topo option. For example, the following command starts a simulated network with single switch and four hosts:

$ sudo mn --topo single,4

The following command starts a simulated network with three switches connected in a line, and each has one host:

$ sudo mn --topo linear,3

You can use the following command to show more options of starting Mininet:

$ sudo mn -h

After starting Mininet CLI, at the mininet> prompt, type the following command to show all the nodes on the simulated network:

mininet> nodes

The following command shows network information on the simulated network:

mininet> net

The following command shows the information of all the links on the simulated network:

mininet> links

The following command shows more detailed information of all the nodes on the simulated network, including the IP address of each node:

mininet> dump

The following command shows the help information of all the Mininet CLI commands:

mininet> help

Figure 5.21 shows the output of some of the preceding commands in the Mininet virtual machine.

The ifconfig command is a useful Linux command to display network information and to configure a network interface. It is similar to the Windows ipconfig command you used earlier. You can run the ifconfig command on the h1 node by typing the following command to display the network information for the h1 node:

mininet> h1 ifconfig

In this case, all the nodes use 10.0.0.0 IP addresses. Your simulated network might use different IP addresses, such as 192.168.0.0, depending on your computer network settings.

Similarly, the ps Linux command displays information about the processes running on the system. The ps -a command displays all the processes. You can run the ps -a command on the h1 node by typing the following command:

mininet> h1 ps -a

The ping Linux command can check the connections between two nodes. You can ping from host 1 to host 2 by typing the following command. The ping -c 10 command means to ping 10 times. Figure 5.22 shows the ping results.

mininet> h1 ping -c 10 h2

You can also ping all nodes by typing the following command:

mininet> pingall

You can test the TCP bandwidth between the h1 node and the h2 node by typing the following command:

mininet> iperf

You can start the x terminal on the h1 node by typing the following command:

mininet> xterm h1

You can start a simple Python web server on the h1 node by typing the following command. The 80 means run the server on port 80, and & means to run the server in the background.

mininet> h1 python -m SimpleHTTPServer 80 &

You can view the web server content in text mode on the h2 node by typing the following command:

mininet> h2 wget -O - h1

You stop the Python web server on the h1 node by typing the following command:

mininet> h1 kill %python

You can also add new nodes to the network. The following are the steps. Type the following command to add a node called h3 to the network.

mininet> py net.addHost('h3')

Type the following command to add a new link between the s1 node and the h3 node:

mininet> py net.addLink(s1, net.get('h3'))

Type the following command to attach a new Ethernet interface on the s1 node:

mininet> py net.attach('s1-eth3')

Type the following command to configure the IP address for the h3 node:

mininet> py net.get('h3').cmd('ifconfig h3-eth0 10.0.0.3')

Type the following command to display the network information on the h3 node:

mininet> h3 ifconfig

Figure 5.23 shows the output of the preceding commands. As you can see, you should also be able to ping the h3 node from other nodes.

You can visualize the Mininet topology by using the online Mininet Topology Visualizer tool at the following web site:

http://demo.spear.narmox.com/app/?apiurl=demo#!/mininet

Just copy and paste the mininet> links results and the mininet> dump results into the page, and click the Render Graph button to generate the network topology, as shown in Figure 5.24.

Finally, you can exit the Mininet simulation by typing the following command:

mininet> exit

You can clean up the Mininet simulation by typing the following command:

$ sudo mn -c

For more information and advanced usage about Mininet OpenFlow, such as configuring the routers and switches, as well as using an Xming server, Putty.exe, and more, please visit the following resources:

https://github.com/mininet/openflow-tutorial/wiki

http://mininet.org/

https://github.com/mininet/mininet/wiki/Introduction-to-Mininet

https://academy.gns3.com/p/sdn-and-openflow-introduction

https://www.cisco.com/c/en_uk/solutions/software-defined-networking/overview.html

5.7.2 Getting Started with Floodlight

The following is the Project Floodlight web site. It has a detailed “get started” guide for instructions on how to download and install Floodlight.

http://www.projectfloodlight.org/getting-started/

5.7.3 Getting Started with OpenDaylight

The following is the wiki site of OpenDaylight. It shows how to download the OpenDaylight software and shows how to get started by setting up the development environment and getting the code.

https://wiki.opendaylight.org/view/Main_Page