VELNET (Virtual Environment for Learning Networking) Bruce Kneale, Ain Y. De Horta, Ilona Box School of Computing and Information Technology University of Western Sydney, AUSTRALIA
[email protected] [email protected] [email protected]
Abstract The problems of providing a real, physical specialist laboratory to teach computer networking such as, the lack of funding and physical space and the risks and threats to the network environment and infrastructure, can be solved by the use of a virtual learning environment. Velnet is such a virtual learning environment that we have developed and used successfully. Velnet consists of one or more host machines and operating systems, commercial virtual machine software, virtual machines and their operating systems, a virtual network connecting the virtual machines, and remote desktop display software. In order to be able to present more computernetworking concepts and to improve on our original version of Velnet we have been developing a virtual reality overlay. This virtual reality overlay allows students to build virtual networking topologies in a virtual lab. This paper describes Velnet our virtual environment for learning networking and the virtual reality overlay under development. Keywords: virtual computer networking.
1
learning
environment,
Velnet,
Introduction
The importance of teaching networking in a computing degree is recognised, for instance, the Joint Task Force on Computing Curricula (2001) states “networking…[has] become the underpinning for much of our economy”. The report goes on to state: Mastery of [net-centric computing] involves both theory and practice. Learning experiences that involve hands-on experimentation and analysis are strongly recommended as they reinforce student understanding of concepts and their application to real-world problems. Laboratory experiments should involve data collection and synthesis, empirical modeling, protocol analysis at the source code level, network packet monitoring, software construction,
Copyright © 2004, Australian Computer Society, Inc. This paper appeared at the 6th Australasian Computing Education Conference (ACE2004), Dunedin, New Zealand. Conferences in Research and Practice in Information Technology, Vol. 30. Editors, Raymond Lister and Alison Young. Reproduction for academic, not-for profit purposes permitted provided this text is included.
and evaluation of alternative design models. All of these are important concepts that can best understood by laboratory experimentation. The report also lists learning outcomes, such as, “Install a simple network with two clients and a single server using standard host configuration software tools such as DHCP.” and “Implement a network firewall.”. It is also recognised elsewhere that learning about networking requires a hands on learning experience (AGILENT TECHNOLOGIES, 2001, Asia Computer Weekly, 1999, Costa, 2000, Kasim, 2002, Logical UK Ltd, 2002, Mateyaschuk, 1999, Montgomery, 2002, NORTEL NETWORKS, 2002). However, providing hands on practical learning experiences can be problematic. The problems include: 1) lack of funds to establish and maintain networking hardware and software in sufficient quantity so that each student can have equal access to the learning environment inside and outside scheduled classes; 2) lack of physical space, both air-conditioned space for servers, and for the hardware used directly by the students; 3) lack of a secure network environment that can reliably protect the other services provided on the same machines or network; and 4) labs are often needed for multiple subjects with diverse software and hardware configurations. Fortunately, “networking technology has become an essential pedagogical tool” (The Joint Task Force on Computing Curricula et al., 2001). We are developing a network accessible virtual learning environment to address these problems. In this paper we describe our secure, cost-effective approach to hands-on networking using a virtual learning environment we call Velnet (Virtual environment for learning networking) (Kneale and Box, 2003). We have enhanced Velnet with a virtual reality overlay. Velnet was developed and successfully piloted in 2003. Velnet has overcome some of our problems with providing an authentic learning environment. Velnet enabled us to present computer networking concepts about configuring devices. Our development of a virtual reality overlay for Velnet enables us to present computer-networking concepts about physical and logical topology as well as configuring devices.
2
Velnet (Virtual environment for learning networking)
Velnet uses existing hardware and software to present a virtual network environment in which students can securely and safely learn about and experiment with networking technologies. The components of Velnet and their relationships, shown in Figure 1, are: the host machine; the host machine’s operating system; VMWare™, commercial virtual machine software, installed on the host; virtual machines and their operating systems; a virtual network connecting the virtual machines, and; remote desktop display software. Velnet, without the overlay, shows up as several windows on the host machine’s desktop, Figure 2.
2.1
Host Machine
Velnet is designed to operate on a single desktop host computer. As the host computer will be emulating several virtual machines the number of virtual machines that can exist simultaneously on the host computer depends on the amount of memory and CPU speed. The amount of memory (RAM) required on the host machine is calculated based on the memory requirements of the host operating system plus the memory requirements of the virtual operating systems that are will be running concurrently. The required disk capacity on the host machine is calculated based on the storage requirements of each operating system whether or not any virtual machines are running. For example, a configuration of five virtual machines, each requiring 3GB disk space and a host requiring 3GB of disk space would require a minimum total host disk capacity of 18GB. Our first experiment with Velnet had one host computer running Windows® 2000 with five virtual machines. Four of the virtual machines ran Windows® NT 4 Workstation and the fifth virtual machine ran Windows NT® 4 Server that acted as a routing device. The Velnet pilot used six host machines running Windows® XP, each with three virtual machines. One virtual machine ran OpenBSD while the other two virtual machines ran Windows® NT Workstation. Windows® XP remote desktop tools were used to provide students with remote connectivity.
Figure 1. The components and configuration of Velnet
Figure 2. A Windows 2000 host with five virtual machines, each running Windows NT
2.2
Host Operating System (Host OS)
Windows® 2000 was initially chosen as the operating system for the host computer for its ease of configuration and student familiarity with the system. Minor difficulties were encountered with this operating system when using VMWare™ and VNC, as the two applications contended
2.3
VMWare
Velnet requires software that provides multiple hardware environments on the one host computer. Open-source applications that provide virtual PC hardware were not generally available. Two commercial products were reviewed for applicability to this project: VMWare™ (VMWare Workstation, 2002) and Virtual PC™ (Virtual PC, 2002). Both applications provide virtual Intel® x86 hardware environments that are compatible with various host operating systems. While both applications provide virtual hardware environments, the desired network configuration for teaching purposes required connection of two virtual machines via a virtual serial line. At the time of writing, VMWare was the only product that provided the capability to connect two virtual machines via a virtual serial line (VMWare, 2002). Network configurations that do not need serial lines may be acceptable for use with Virtual PC (Connectix, 2002).
2.4
Virtual Machine
The virtual machine emulates an Intel® x86 hardware environment. From the time of “switching on” the machine it appears as if it were a real, physical computer (Figure 3); just that what would initially appear on a monitor, if the machine were real, is now confined to a window on the host machine’s desktop. Desktop of the host machine
for mouse pointer control as the two virtual machines were running GUI based operating systems. Removing VNC and using Windows XP remote desktop access resolved this problem. The problem was not apparent for virtual machines running text based operating systems such as OpenBSD. Even though there is one host machine, each virtual machine has a separate identity, and as such, appropriate licensing must be obtained for any operating system, or other software, used. The choice of Virtual OS for the virtual machines in the initial standalone experiment was Windows® NT Workstation and Windows® NT Server. The choice of virtual OS for the virtual machines for the pilot was OpenBSD and Windows® NT Workstation. Judicious use of open source operating systems such as Linux® and OpenBSD provides less expensive alternatives to proprietary products.
2.6
Virtual Network
The virtual network runs in VMWare, as a function of this application. Some of the components of a physical network such as switches and network adapters are emulated in the virtual network. Routing is supported by operating systems that have this function and are installed on the virtual machines.
2.7
Remote Desktop Display (RDD)
Remote desktop display software allows students to remotely control the desktop of the Velnet host machine from anywhere on the Internet. Most RDD software is client/server type. The server portion runs on the host machine. The client portion of the system runs on any machine from which a student wishes to remotely control the host machine’s desktop. An example of RDD is Remote Desktop for Windows® XP, which is built into the Windows® XP operating system, provides the server portion of the RDD. Microsoft® supplies client software for most Windows® operating systems and certain Apple® operating systems.
Start-up screen of a virtual machine
Using RDD restricts the number of users of Velnet to one at a time per host machine. With multiple host machines this is not a problem, for instance in the pilot six students used the six host machines. However, being able to share host machines, i.e. more than one user at a time, would be better.
3 Figure 3. The start up of a virtual machine appears the same as a physical computer.
2.5
Virtual Operating System (Virtual OS)
The choice of operating system depends on desired learning outcomes, cost, and memory available in the host machine.
Connecting to Velnet using RDD
Remote desktop display software allows students to interact with the virtual network from anywhere on the Internet. For the purposes of our pilot, access was restricted to client computers within the institution’s network. An example of multiple Velnet sessions using RDD is shown in Figure 4. Velnet is running (six separate times) on six host machines. Each Velnet session has an individual, virtual network (Figure 4 a). Students gain remote access via a computer (Figure 4 b) connected to
an institution’s network (Figure 4 d). A network firewall (Figure 4 e) is used both to protect the broader network community from accidental or malicious actions within the virtual network and to restrict access to the virtual network for only those students enrolled in the subject. The firewall is configured to only allow RDD sessions. Any configuration problems either with the virtual networks or the host computers will be confined to the teaching network (c).
Winn and Jackson go further by proposing that though a virtual environment may be less real, such as our virtual reality overlay, the trade-off with cost does not make the outcomes poorer. It may be that skills are more easily transferred because the virtual environment is lower fidelity. A student learning in a high fidelity VE may expect that the real world will be the same and therefore be confused by small changes. Lower fidelity VEs make the student generalise what they have learnt in the real world. The Velnet virtual reality overlay is low fidelity. The overlay is expected to give an idea or feel of working with a number of, what would be in reality, physically separate machines, even though the user would be working in front of one monitor, i.e. one physical position. 2) Virtual environments are safer.
Figure 4. Typical remote access session to Velnet using RDD. The first pilot of Velnet restricted access to client computers within the institution as stated previously. However, once security issues for wider Internet access are resolved, it would be possible for students to connect from anywhere. As mentioned the use of RDD access to Velnet restricts the use to one host machine per student at a time. There are cost benefits to using Velnet with RDD. For instance to teach the learning outcome “Install a simple network with two clients and a single server using standard host configuration software tools such as DHCP.” (The Joint Task Force on Computing Curricula et al., 2001), we need one computer plus software per student, not multiple machines plus software. However, it still requires a significant investment. It would be an enhancement to Velnet if we could run one host accessed by many users. The enhancements to Velnet of running one host accessed by many users lead us to consider how simulations and virtual reality can contribute to virtual learning environments.
4
Virtual Learning Environments and Virtual Reality
Winn and Jackson (1999) presented 14 propositions about the use of virtual reality in learning environments. Those applicable to our enhancement of Velnet include: 1) Virtual environments (VEs) are cheaper than a physical environment. We found that Velnet accessed via RDD does save costs, even with the extra RAM, CPU, and disk space required. If we could have one Velnet host accessed via virtual reality overlays by many users we expect it would still be cheaper.
The VE provides a safe place to practice risky tasks. In our situation the safety of the network is one concern. Another is that the students feel safe to experiment without causing damage for which they may feel they would be reprimanded. This was also argued by Forbus et al. (1999). Velnet on one host server with multiple client virtual reality overlays could be implemented similar to that shown in Figure 4. The differences being that one host (Figure 4 a) would be required not six and the clients (Figure 4 b) would each have a client virtual reality overlay running rather than RDD. 3) Natural interaction with the VE allows students to experience metaphorical concepts and undetectable phenomena. In a real environment, students would apply networking skills and knowledge by using a keyboard, a mouse, and a monitor to work on a device or access one device from another. Velnet emulates this natural interaction. However, if we want the students to see, say a packet be routed through a network, we cannot do this in a real environment. We can do this in a simulation. Taking this idea further, the virtual reality overlay of Velnet plus a VE helmet and glove would put the students in a VE where they could become part of the network from a bit on a bus, to a bridge, server or firewall; time can be slowed; concepts and non-detectable physical forms can take shape. (Winn and Jackson refer to this as reification). 4) There is growing evidence that students, particularly those not academically inclined or committed (such as the majority of students undertaking higher education (Biggs, 1999)), are more likely to do well and develop familiarity with the subject material in a VE. The students also then see a VE as valuable. Providing authentic learning experiences by modifying the laboratory environment may increase student engagement in the learning process and lead to high-order thinking in students (Cruikshank, 2002).
This proposition could be supported or refuted by reflecting on the teachers’ experiences of using Velnet and the virtual reality overlay. For now, we are encouraged by the growing evidence that Velnet ant its overlay will improve the learning experiences of our students. 5) Constructivist concepts of learning and “first-hand” experience (Clancey, 1993) allows students to take what is familiar to them and add to this knowledge.
Virtual Lab Area
Component Libraries
In this case the students will be using a GUI, something that is familiar, and adding knowledge about networking that they can gain by experiencing the VE. The VE can then be used to assess the student in an activity where the student demonstrates their construction of meaning given a problem to solve. Such an authentic assessment being judged on the student’s holistic performance of the tasks (Montgomery, 2002). This proposition is also lent weight by the advice from Braathen and Robles (2000).
Figure 5. Velnet Virtual Reality Overlay. Figure 6 shows all the current component libraries and components within each library. The number of components and/or libraries can be expanded.
6) The VE situates learning in a real context. The appearance of each virtual machine in Velnet is as it would appear as if a separate physical machine. Velnet has presence; the students can be convinced that they are working with a real network. They would directly connect their experience with Velnet to the real situation. Velnet is “real” enough for common networking tools such as protocol analysers and network management software to be employed to solve real problems. Students will encounter the same challenges using Velnet as in the real world: idiosyncratic operating systems, buggy implementations of network software, and difficult to understand user documentation. Of the 14 propositions presented by Winn and Jackson (1999) we found the six discussed above supportive of enhancing Velnet with a virtual reality overlay.
5
Velnet Virtual Reality Overlay
We enhanced the functionality of Velnet by developing a virtual reality overlay. The overlay is designed to help students to learn more about networking by allowing them to construct and experiment with different networking architectures in a virtual lab environment. The overlay lets the students “see” the network as machines in a topology, rather than as windows on a host machine such as the case with “raw” Velnet, shown in Figure 2. The overlay is a safe, low fidelity VE. The virtual reality overlay is intended as client software. Velnet is to run on a server. The interface to the virtual reality overlay is shown in Figure 5. It has two main panes, within the interface: 1) on the left, a virtual lab area where networking components can be placed, interconnected and configured to form a networking architecture; and 2) on the right, component libraries containing various types of networking and networking related components.
(a)
(b)
(c)
(d)
Figure 6. Component Libraries: (a) Network Nodes; (b) Interfaces; (c) Media; (d) Network Devices.
5.1
Using the Velnet Virtual Reality Overlay
The construction of a network using the Velnet overlay is a simulation of doing the same thing in the real world. A user of the virtual reality overlay can construct a network by using familiar GUI interactions. A student would select a component as needed from the libraries on the right, drag and drop the component into the virtual lab area on the left. Once all the components have been placed, the virtual lab area would be appear as a metaphor of the equivalent real world network architecture. The student would then configure the devices in the network, in the same way they would do the configuration in the real world. Then they would be able to test their construction. As with the real world, the components and their configuration would have to be correct for the virtual network to work. Sections 5.1.1 to 5.1.3 describe a worked example using the Velnet virtual reality overlay.
5.1.1
Constructing the Network Architecture
Consider, as an example, the network architecture shown in Figure 7, it is typical of an exercise that is given to students.
Figure 7. Typical network architecture By using the virtual reality overlay of Velnet it is possible for students to construct the architecture shown in Figure 7 and to see their progress as they work. The steps to construct the architecture shown in Figure 7 are: 1. Select the library that contains a router, shown in Figure 6d. The student would need to know that a router is a network device in order to select the correct library, or hunt through the libraries to find it. 2. Select and place the router icon by clicking and dragging it into the virtual lab area. This completes the placement of the component in the virtual lab area. This is the same as placing a router on a workbench so that it can be configured for a particular network. 3. Complete the dialogue box requesting details about the device. For workstations, routers, and servers a name is requested. For repeaters, the user is given a selection of available virtual networks from which they select one. 4. Install a network interface card by dragging the appropriate icon from the library (Figure 6 b) and “install” it in the router by dropping it over the router icon in the virtual lab area. In order to connect a router to a network, it needs a network interface card, just as in the real world. Step 4 imitates the real world installation procedure. The network interface icon changes to a red tick mark on the router, workstation, or repeater. Steps 1 to 4 are repeated for each workstation, repeater, and router in Figure 7. 5. Once all the components are in place they will need to be connected by media. This is accomplished in a similar way as other components, selecting from a library and dropping them in place. The media is anchored at a start point, the interface tick mark on a component, by a mouse click, and then tied off at another component’s interface, by another mouse click. The architecture shown in Figure 7 would appear in the virtual lab area as shown in Figure 8.
Figure 8. The architecture as shown in Figure 7 constructed in the virtual lab area. As in the real world, once the networking media connects all of the components of the network to form a network architecture, it is then necessary to configure components such as routers, servers, and workstations. In the example in Figure 8, the components that require configuration are the routers and workstations.
5.1.2
Configuring a Router
In the real world, to configure a router requires us to firstly power up the router. In the virtual lab this is also the first step when configuring. This powering up process would appear as shown in Figure 9. The router is selected and started by selecting "Start Virtual Machine" in the popup menu. This starts the virtual router, which in this case is a virtual machine running OpenBSD setup to be a router. Once the router is powered up we can open a console and perform the required configuration. Once again select the router and from the popup menu select "Console" as shown in Figure 10. The powering up would connect the client software to the server running Velnet. The configuring would be a capture of a window presented by VMWare.
5.1.3
Configuring a Workstation
The process of starting and configuring a workstation is identical to that of configuring a router, except that when we open a console we are now opening a virtual machine running Windows NT 4 (WS) as seen in Figure 11. We now have a fully functioning virtual network that can be tested and used to explore networking and networking related issues.
6
Future Enhancements and Direction
The first attempt at the virtual reality overlay runs on top of the current incarnation of Velnet and is therefore constrained by current limitations. A project to produce a client/server version of Velnet and its virtual reality overlay is underway. Once the client/server version is complete (anticipated by the end of 2003) the next step in the Velnet project is to
access to Velnet via the Internet, within the institutions available laboratories at first and then the world wide web. There are arguments for and against 24 hours a day, seven days a week access (O'Donoghue et al., 2001, Piccoli et al., 2001). However, the against arguments can be won over with careful instructional and assessment design. The benefits expected and, arguably, the inevitability of the role the Internet will play in course delivery, make completion of this phase of the project probable in the next year. In the furture we are planning, as part of the instructional and assessment design, multimedia support material, such as video of the real world activities, like installing a network interface card (Step 4 in section 5.1.1). These instructional support materials are another facet of the Velnet project. With adequate resources on the host machine it is possible that Velnet could be used for more than learning about networking. It is possible that distributed computing concepts could be presented using Velnet. Even, with a host machine powerful enough, simulating a load test of a client/server application, written, installed, and maintained by students.
Figure 9. Starting the router.
Another possible long-term direction for Velnet may be in the ability to convert the virtual reality overlay into an articulate virtual laboratory (AVL) (Forbus et al., 1999). An AVL, according to Forbus et al. would address problems of marking, timely feedback, coaching, and guidance.
7
Summary
Velnet is an ongoing project to develop a feasible, reliable, authentic learning environment in response to the growing need for net-centric computing in computing degrees. We have progressed from “raw” Velnet to a virtual reality overlay of “raw” Velnet. Currently, progress is being made on a client/server version of Velnet and its overlay.
8 Figure 10. Opening a router console.
Figure 11. Opening a Workstation console. test the provision of 24 hours a day, seven days a week
References
Agilent Technologies (2001): Agilent technologies: Agilent technologies offers free network troubleshooting software to colleges and universities; professional tool helps teach students modern communication protocols and networking principles. M2 Presswire:1. Asia Computer Weekly (1999): Networld+interop99(n+i99) preview: Sign up for the atm classroom at fore university. Asia Computer Weekly:1. Biggs, J. (1999): Teaching for quality learning at university. Buckingham, England, Society for Research into Higher Education and Open University. Braathen, S. and Robles, M. (2000): The importance of assessment in business education. National Business Education Yearbook, 2000:11-24.
Clancey, W. J. (1993): Situated action: A neuropsychological interpretation: Response to vera and simon. Cognitive Science, 17:87-116. Connectix (2002): Virtual pc for windows user guide. Connectix Available at. Accessed on October 2002. Costa, G. (2000): The cisco certification programs. Exploring Windows NT, 8(4):14. Cruikshank, B. J. (2002): Can problem-based instruction stimulate higher order thinking? Journal of College Science Teaching [H.W. Wilson - EDUC], 31(6):374. Forbus, K. D., Whalley, P. B., Everett, J. O., Ureel, L., Brokowski, M., Baher, J. and Kuehne, S. E. (1999): Cyclepad: An articulate virtual laboratory for engineering thermodynamics. Artificial Intelligence [H.W. Wilson - AST], 114(1/2):297. Kasim, S. (2002): More schools to offer ict by 2005. Computimes Malaysia:1. Kneale, B. and Box, I. (2003): A virtual learning environment for real-world networking, Informing Science, Pori, Finland, pp:71, Logical Uk Ltd (2002): Logical uk ltd: Sheffield hallam university awards network upgrade to logical. M2 Presswire:1. Mateyaschuk, J. (1999): Certification checks out for it. Computer Reseller News, (846):52. Montgomery, K. (2002): Authentic tasks and rubrics: Going beyond traditional assessments in college teaching. College Teaching, 50(1):34. Nortel Networks (2002): Nortel networks: University of miami to deploy next generation metro optical solution from nortel networks; faster, more cost-effective transport network expected to enhance academic research. M2 Presswire:1. O'donoghue, J., Singh, G. and Dorward, L. (2001): Virtual education in universities: A technological imperative. British Journal of Educational Psychology, 32(5):511-23. Piccoli, G., Rami, A. and Blake, I. (2001): Web-based virtual learnign environments: A research framework and a preliminary assessment of effectiveness in basic it skills training. MIS Quarterly, 25(4):401-426. The Joint Task Force on Computing Curricula, Ieee Computer Society and Association for Computing Machinery (2001): Computing curricula 2001: Computer science. ACM Journal of Educational Resources in Computing, 1(3). Virtual Pc (2002): Connectix Corporation, San Mateo, California, USA. Vmware (2002): Vmware workstation 3.2 user's manual. VMWare Inc. Available at http://vmwaresvca.www.conxion.com/software/ws32_manual.pdf. Accessed on October, 2002 2002. Vmware Workstation (2002): VMWare Inc., Palo Alto, California, USA. Winn, W. D. and Jackson, R. (1999): Fourteen propositions about educational uses of virtual reality. Educational Technology [H.W. Wilson - EDUC], 39(4):5-14.
