Computing Labs on Virtual Environments: A Flexible, Portable, and Multidisciplinary Model

Depósito de Investigación de la Universidad de Sevilla https://idus.us.es/ This is an Accepted Manuscript of an article published by IEEE in: J. Muñoz-Calle, F. J. Fernández-Jiménez, T. Ariza, A. J. Sierra and J. M. Vozmediano, "Computing Labs on Virtual Environments: A Flexible, Portable, and Multidisciplinary Model," in IEEE Revista Iberoamericana de Tecnologias del Aprendizaje, vol. 11, no. 4, pp. 235-241, Nov. 2016, doi: 10.1109/RITA.2016.2619139. “© 2016 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other Works”  1 Title— Computer labs on virtual environments: A flexible, 2 portable and multidisciplinary model. 3 4 Abstract— Computer-aided teaching of practical subjects in 5 engineering education poses new challenges. Computers have 6 to be configured to support the particular requirements of each 7 subject. Virtual Environments allows the building of a Virtual 8 Machine (VM) tailored to the requirements of each subject, 9 allowing flexible, versatile and low cost laboratory 10 configurations. However, the use of multiple VM at a shared 11 computing facility poses new problems on the lab sessions, both 12 technical and performance related. To solve these challenges, 13 we have developed the Virtual Machine on-Demand (VMD) 14 tool. This tool automates the adaptation of each VM to the 15 environment in which it is deployed. As a result, the creation of 16 computer-aided engineering laboratories on a shared 17 computing facility using VM is possible and profitable. 18 19 Index Terms — Engineering education computer-aided 20 laboratories; Virtual environments; Shared computing facility; 21 Computer applications; Programming languages. 22 23 I. INTRODUCTION 24 he University of Seville [1] is offering the Degree on 25 Telecommunication Technology Engineering (GITT 26 from its Spanish name) [2] since 2010, according to the 27 European Higher Education Area (EHEA) agreements [3-4]. 28 The EHEA agreements aim to promote European 29 convergence in education. So far, one of its most significant 30 changes, related to traditional teaching, is to focus learning 31 on the students effort, thus increasing the amount of subjects 32 with computer-aided laboratory practices. Supporting that 33 high number of practical subjects brings new challenges. 34 Strict budget limitations mandate the sharing of computer 35 resources for teaching, despite the different and even 36 mutually incompatible software requirements, of different 37 subjects. Computers are located at a shared computing 38 facility, which in our case is named Computer Center (CC). 39 There, the staff has to maintain an increasing number of 40 computers configured for any subject. To further complicate 41 their job, the computer park is quite heterogeneous, due to 42 technological changes and to the stepped acquisition and 43 replacement procedures. 44 Traditionally, that heterogeneity in both hardware and 45 software requirements of subjects has been managed by the 46 CC with the aid of a cloning tool, such as the IBM’s 47 The authors are with the Department of Telematic Engineering, University of Seville, Spain (phone: +34 954 48 73 85; e-mail: {javi, fjfj, matere, antonio, jvt}@trajano.us.es). DOI (Digital Object Identifier) Pending Rembo™ tool [5]. However, the increased number of 48 practical subjects brought by the EHEA led to a new 49 scenario, with an excessive number of different Operating 50 Systems (OS) or lab classes to be accommodated in the CC 51 by this approach. 52 To address the above difficulties, a growing number of 53 professors decided to use Virtual Machines (VMs) in their 54 subjects. The VMs allow flexible, versatile and low cost 55 computer laboratories [6-10]. Despite the excellent results of 56 these Virtual Environments (VE), it was detected that its use 57 generated two new groups of shortcomings. On the one 58 hand, hardware heterogeneity and CC security demands 59 required to adapt the configuration of each VM to the every 60 specific computer type, which caused difficulties from a 61 technical point of view. On the other hand, many students 62 found it hard to use the native user interface to play VMs (in 63 our case, VMware Player [11]), reducing the educational 64 performance of the class sessions. 65 The aim of this work was to solve these two groups of 66 problems. We present a model based in the Virtual Machine 67 on Demand (VMD) tool. This model makes the flexible and 68 efficient use of adapted virtual machines in shared computer 69 labs possible. It has been successfully applied for some 70 years at our CC, supporting multiple subjects from our 71 Department of Telematic Engineering. VMD allows the 72 student to bring their tailored VMs on removable media, and 73 to play them easily in any computer, being automatically 74 adapted to the specific requirements of computer and CC 75 being used. 76 The rest of this article is organized as follows. The next 77 section introduces the reader to virtualization techniques. 78 Section III presents the technical and educational context 79 that motivated the development of the VMD model. Section 80 IV describes the model and functionality offered by VMD, 81 while Section V provides its detailed operational 82 description. Next, Section VI summarizes the main benefits 83 of using this model, corroborated by the data contained in 84 Section VII. Section VIII shows several improvements that 85 are currently being applied to VMD. Finally, Section IX 86 concludes the paper. 87 II. VIRTUALIZATION BACKGROUND 88 In computer context Virtualization is the construction of a 89 software version, called Virtual Machine (VM), of an 90 existing resource called real machine. Fernández and 91 colleagues [12] describe the different types of virtualization, 92 what a virtual machine is and different experiences carried 93 out in education with virtual machines. Among different 94 types of Virtualization, hardware virtualization is focused on 95 Computing Labs on Virtual Environments: a Flexible, Portable and Multidisciplinary Model Javier Muñoz-Calle, Francisco J. Fernández-Jiménez, Teresa Ariza, Antonio J. Sierra and Juan M. Vozmediano T creating a virtual machine that acts like a real computer with 96 an operating system (OS). 97 The use of virtual machines in the engineering education 98 laboratories of the Department of Telematic Engineering of 99 the University of Seville, discussed in [12], is based on type- 100 2 hypervisors: virtualization software runs on the host 101 operating system as an application. The guest operating 102 system that runs on top of virtualization software usually 103 requires modifications to run in this virtual environment. 104 Fig. 1 shows the structure of a system with this kind of 105 virtualization, used by the developed VMD tool. 106 III. TECHNICAL AND EDUCATIONAL CONTEXT 107 This section looks into the difficulties that arose with the 108 evolution of the syllabus in more depth, and explains how 109 the VMD model offers an excellent solution for them. 110 111 A) Background before EHEA 112 The GITT studies of the School of Engineering at the 113 University of Seville started in the 2010-2011 academic year 114 [13]. The syllabus is divided into 240 credits, spread over 115 four years. 116 One of the requirements for the design of the syllabus was 117 to strengthen the practical nature of teaching as much as 118 possible, a requirement that has been intensively conducted 119 in the subjects taught by the Department of Telematic 120 Engineering. The involvement of the Department in the 121 delivery of the syllabus is summarized in [4]. About 50% of 122 the offered credits are practical ones, and only one out of 20 123 subjects does not have any laboratory sessions. 124 To support these practical credits, each subject requires 125 the use of a computer lab, offered by a shared Computer 126 Center (CC) facility,. The CC provides 385 computers in 11 127 classrooms. Different software requirements between 128 subjects (e.g. different OS or mutually incompatible 129 configurations) and hardware heterogeneity have been 130 traditionally managed with the aid of a cloning tool such as 131 the IBM’s Rembo™ tool. This allows booting from different 132 operating systems and configurations as needed for the 133 several subjects, as well as a fresh-boot for each computer 134 (always booting from the same state), which eliminates 135 possible manipulations from the previous class sessions. 136 137 B) Scenery after EHEA. Solution: VM 138 Until the appearance of EHEA, the aforementioned 139 cloning tools were enough to support computer practices of 140 different subjects. However, the increased number of 141 practical subjects under the new educational scenario 142 brought by the EHEA led to the emergence of new problems 143 that these tools cannot solve: 144 -- Firstly, the required number of different OS for all 145 these subjects is excessive, and can not be installed 146 natively on each computer due to disk size constraints. 147 The cloning tool could perform a complete restore of the 148 required OS in each boot (sharing the same partition 149 among multiple OS), but that would lead to an 150 unacceptable booting time at the beginning of each 151 session. Also, the maintenance of so many OS versions 152 would require an excessive maintenance work from the 153 CC staff. 154 -- Secondly, under the new educational scenario of the 155 EHEA, there are subjects that teach OS-administration 156 concepts, which obviously require OS administration 157 privileges. However, the CC policy is to forbid that these 158 administrative privileges are granted to the students to 159 avoid misconfigurations and computer misuse. This 160 makes the computers of the CC unsuitable for subjects 161 that need to modify the computer configuration. 162 -- Finally, the excessive number of practical classes 163 exceeds the capacity of the CC to accommodate them. 164 This represents a major inconvenience that needs to be 165 solved as a priority. 166 167 To solve the above drawbacks we decided to use Virtual 168 Environments (VEs). As stated by [6-10], virtualization 169 techniques adapt to different computer hardware and 170 subjects’ requirements. The use of VEs solved these 171 problems: 172 -- Regarding the excessive number of OS, VEs allows 173 each computer in the CC to have physically installed only 174 one or two host OSs (which can still be managed with a 175 cloning tool), and on these ones VMs are used. This 176 greatly reduces the maintenance work of the computers in 177 the CC, while still allowing a fast booting at the 178 beginning of each session. 179 -- VEs allows the granting of administrative privileges 180 to the students without putting the host OS at risk 181 -- The use of VEs allows the student to replicate the 182 laboratory environment in his own computer [6-10]. The 183 students can use their own laptops in the CC alleviating 184 the problem of the insufficient number of computers in 185 labs. Besides, the students may continue their practical 186 learning at home. 187 188 C) VM problems. Solution: VMD model 189 According to the above, VEs allowed flexible, versatile 190 and low cost laboratories to teach practical subjects in 191 engineering. Based on these advantages, in recent years a 192 growing number of professors of the GITT began using 193 VMs to provide the suitable software environment for their 194 labs. However, during this time we have identified a number 195 of shortcomings when multiple VMs are provisioned for 196 different subjects at a CC. 197 Even though it is desirable that a single VM would be 198 valid for several subjects, each one has its own software 199 requirements, often incompatible with other subjects. For 200 example, two subjects might require a different version or a 201 particular configuration of the OS. Therefore, each subject 202 or group of subjects would use their own VM with the 203 VM1 Host OS Hardware … Apps Type 2 hypervisor Guest OS VMn Guest OS … Fig. 1. System architecture with type 2 hypervisor necessary requirements. Thus, the total number of different 204 VMs to use in the CC may be high. It could be envisioned 205 that the CC staff would install the VMs on the computer’s 206 hard disk using a cloning tool, freeing students from 207 carrying the VM in removable media. However, this choice 208 involves multiple disadvantages: 209 -- When the number of different VMs is very high, it 210 may be physically impossible to place them 211 simultaneously on the hard disk of computers due to 212 space reasons. The use of a network server to download 213 each VM could cause excessive delays in subjects with a 214 large number of students. For instance, in one of the 215 subjects we use a VM that, when compressed, occupies 216 1.02 GB, and the practical classes are hold on several 217 sessions a week. In the academic year 2013/2014 we 218 peaked 86 simultaneous students a session. If the VM 219 were hosted on a network server, 99.96 GB of data would 220 have been transferred at the start of the class. This 221 requires some 15 minutes with the available 1 Gb/s links. 222 As the CC facility is shared, it would also affect students 223 of other subject-s who were working at that time in the 224 CC. 225 -- The cloning tools, such as the Rembo tool used by 226 the CC, do not work well with large files such as those 227 used as virtual hard disks in the VMs. 228 -- For organizational reasons, the CC restricts changing 229 the environment in the PCs to the beginning of the 230 semester. This is especially important in the 231 implementation of the new subjects as they would have to 232 program the final environment at the beginning of the 233 semester. Additionally, this requirement cannot be 234 fulfilled in some subjects, such as those ones that use 235 incompatible software tools that must be installed 236 progressively as the subject is imparted. 237 238 Therefore, the only valid approach is that each student 239 attends to the lab sessions with their VMs on a removable 240 media (USB, DVD, etc.). This requires solving two new 241 problems: 242 1) Technical challenges: On the one hand, computer 243 hardware of the CC is quite heterogeneous , so the 244 hardware configuration of each VM needs to adapt to 245 the physical hardware features of the computer where 246 it will be used (e.g., available physical memory, 247 number of processor cores or number of physical 248 network cards). The creation of a different version of 249 VM for each type of hardware would result in an 250 excessive workload, not being an acceptable 251 alternative. 252 On the other hand, for security and organizational 253 reasons, the CC requires that each VM has a 254 particular network configuration according to the 255 computer that is to be used. For example, virtual 256 network cards in each VM must have a different and 257 known Ethernet address, both to allow dynamic IP 258 address management and to allow physical location 259 of each VM. This problem cannot be solved by 260 creating multiple versions of each VM, hence calling 261 for a necessary solution. 262 2) Negative effects on the performance of the class 263 sessions: We observed that many students, mainly 264 freshmen, had trouble using the native interface 265 provided by the virtualization software (e.g. 266 VMWare Player) to locate and play the VMs, which 267 caused significant delays in the start of each class 268 session. It was also identified that the handling of 269 many different VMs required an extra management 270 effort from the students, distracting them from the 271 educational target of the session. 272 273 The following sections detail the model and operation of 274 the tool VMD developed to solve these issues. 275 IV. VMD FUNCTIONAL MODEL 276 The VMD tool allows using any computer in a CC to 277 develop practices and projects required in any subject, using 278 its custom VM. The teacher of each subject must provide the 279 customized VM, with the proper software applications. The 280 VM is packed in a single compressed file, simplifying the 281 students management tasks. Students will get and store these 282 VMs them in any removable media such as a USB memory 283 or DVD. 284 Each computer of the CC is provided with the VMD tool, 285 which can be easily launched from a desktop icon or a 286 simple command. After running the VMD, the students can 287 play transparently each VM, be it either installed in the 288 computer or saved in a removable media, as shown in Fig. 2. 289 To ensure an optimum portability between different OS 290 and to reduce the requirements for its use, the VMD tool has 291 been developed on POSIX Shell Command Language [13]. 292 VMD provides a command line interface, allowing its use 293 on computers without a graphical environment (e.g., in 294 server administration practices), and offering a comfortable 295 and intuitive user interface that guides the user step-by-step 296 through the different operations (detailed in the following 297 section). Originally, it uses VMWare Player [11] as 298 virtualization environment given its wide use among 299 students. 300 V. DETAILED VMD OPERATIONAL DESCRIPTION 301 VMD allows to automatically locate VMs in removable 302 media and to deploy them in a hard disk folder. It also 303 permits the execution of the deployed VMs. Fig 3 shows a 304 simplified flow diagram of the operation of the VMD. We 305 have omitted some checking user-entered data and error- 306 handling tasks , which do not add relevant information to the 307 explanation. The following describes the steps performed by 308 the VMD: 309 1) Searches for VMs that are already installed on the 310 hard disk of the computer (Fig. 3 - 1). All VMs are 311 deployed under a known directory. This allows for a 312 Removable Media /Internal Disk VMoD VMoD VMoD VMoD VM Deploy & Play VM Selected VM Student Fig. 2. VMD’s scenario list of all installed VMs. 313 2) Displays a menu with the VMs found in step 1, an 314 option to deploy a new VM and another option to 315 cancel the process (Fig. 3 - 2). 316 3) The user input is checked (Fig. 3-3). If the 317 deployment of a new VM is chosen, the proper task is 318 applied. This is detailed below (Fig- 3 - 4). It is only 319 necessary to deploy a VM once. For this reason the 320 tool allows the user to start a VM that was deployed 321 before. 322 4) If in steps 2 or 3 the user selects a VM (Fig. 3 - 5), 323 this VM will be played (Fig 3 - 6). Otherwise the user 324 is prompted to restart the VMD (Fig. 3 - 7). 325 5) If the program is restarted, it goes back to step 1. 326 Otherwise, the program is terminated (Fig. 3 - 8). 327 328 The process of deploying a new VM (Fig. 3 - 4) is a 329 subtask composed of the following steps: 330 -- First, existing VMs are searched in the root directory 331 of any removable media. They are expected to be 332 compressed into a single file. Since the user can have 333 multiple VMs the user is asked to enter a search pattern 334 (Fig. 4 - 4a). 335 -- Then, the VMs are searched in the distinct removable 336 media (Fig. 4 - 4b). 337 -- If no VMs are found (Fig. 4 - 4c), the subtask ends. 338 Otherwise a menu is shown allowing the selection of a 339 found VM (Fig. 4- 4d). Users can also cancel the process 340 (Fig. 4- 4e). 341 -- The selected VM is installed, provided it was not 342 before. (Fig. 4 - 4g/4h). This new VM is marked as 343 selected by the user to be played. 344 -- If the VM were already installed (Fig. 4 - 4f) the user 345 is prompted to delete the installed copy and redeploy it 346 from the fresh copy (Fig. 4 - 4i). According to the user’s 347 answer (Fig. 4 - 4j), if he does not want to delete the 348 existing VM, the process continues as if this VM had been 349 selected in (Fig. 3 - 2). Otherwise, all the files associated 350 with that VM are deleted from the hard disk (Fig. 4 - 4k) 351 and the VM is installed from the removable media (Fig. 4 352 - 4g/4h). In both cases (and provided that no error occurs), 353 this VM is marked as selected by the user to be played 354 (Fig. 3 - 6). 355 356 As noted above, the installation of a new VM from the 357 removable media consists of two steps (Fig. 4 - 4g and 4- 358 4h). In the first step (Fig. 4 - 4g) the VM is deployed on the 359 hard drive of the host. The compressed file is decompressed 360 and verified. 361 In the second step (Fig. 4 - 4h) the VMD adapts the VM 362 configuration to the specific features of the computer. This 363 adaptation consists of two parts: 364 -- On the one hand, the configuration of the VM is 365 modified to suit the physical characteristics of the 366 equipment, allowing the heterogeneous computer 367 hardware of the diverse CC equipments to be transparent 368 to teachers, students and maintenance personal. For 369 example, it adjusts the memory used by the VM to the 370 physical memory available on the computer, the number 371 of processor cores used by the VM to the cores available 372 on the computer’s microprocessor and it adjusts the 373 configuration of the virtual network cards (NAT or Bridge 374 mode, in VMware notation [11]) according to the number 375 of physical network cards. 376 377 -- On the other hand, the VMD adjusts the Ethernet 378 addresses of the virtual network cards with direct network 379 access (also known as cards in bridge mode). These 380 Ethernet addreses are computed from the physical network 381 cards’ addresses of the host computer, where the VM is to 382 be run. This is an organizative and security requirement 383 from the CC: to locate and identify every computer, they 384 must use a unique and known Ethernet address, so that the 385 same dynamical IP address can be be assigned to the same 386 computer. 387 388 This adaptation process is a fundamental contribution to 389 the VMD, and permits the creation of computer aided 390 engineering laboratories on a shared computing facility in an 391 efficient way. 392 Fig. 3. VMD’s operational description 393 394 395 VI. VMD BENEFITS 396 The proposed VMD model permits the teaching of 397 subjects with computer laboratories at the CC using VMs in 398 a fast and guided way. In more detail, the advantages of 399 using this model are: 400 1) It allows manipulating the configuration of the VMs 401 when they are deployed. The configuration is adapted 402 to the particular hardware and network of the CC. 403 Many CCs, such as our CC, demand that the VMs 404 fulfill certain settings to meet the organizational and 405 security requirements. 406 2) It improves the performance of the class sessions: as 407 opposed to other virtualization programs, VMD does 408 not burden the student with finding the proper VMs 409 in the directory structure of the system. In addition, 410 VMD offers a guided user interface far simpler than 411 those offered by the virtualization software tools and 412 therefore minimizes the initial start-up time at each 413 session and the student lab-management effort. 414 3) It solves the problem of deploying the VMs on the 415 computers of the CC. The VMs are not deployed 416 using cloning tools, but students will attend the CC 417 with the VM stored in a removable media and VMD 418 will deploy it on the hard disk. 419 4) After deploying the VMs, they are played from the 420 hard drive and not on the removable media such as an 421 USB flash memory drive. This provides: 422 -- Faster working speed (hard drives are usually 423 faster than USB flash memory drives). 424 -- Better aging of removable USB drives, as 425 they are read and not written (the flash memories 426 have a maximum number of rewrites). 427 -- More flexibility, as the VMs can be carried 428 on other memory devices such as DVD-R (does 429 not permit writing), in addition to USB drives. 430 5) Virtualization software tools store each VM as a set 431 of files. However, VMD stores each VM as a single 432 compressed file. The advantages of this approach can 433 be summarized as follows: 434 -- Faster copy to the hard disk 435 -- Less space used on removable media 436 -- Students can download the VMs from the 437 Internet and use them directly without the need to 438 decompress them manually. 439 -- The distribution of each VM in a single file is 440 more comfortable than in a folder (each subject 441 can use its own VM, therefore students may have 442 to use a lot of VM) and avoid possible errors 443 (ensuring that the VM is complete, not lacking 444 files that students have forgotten to copy). 445 446 All these advantages motivated the development of this 447 model, both to ease the students and professors’ tasks as 448 well as permitting the VMs use in the CC. 449 VII. VALIDATION 450 Since the academic course 2012/2013, the VMD model 451 has been used by more than 1000 students in the subjects of 452 the Department of Telematic Engineering listed in the table 453 I, about a CC equipped with more than 300 computers. That 454 allowed us to prove that this model resolves the two types of 455 problems that led to its development: 456 -- The possibility to use a VM in a CC independently 457 of the cloning mechanism that is used, the hardware 458 characteristic of the equipment and the network 459 requirements that it may demand. 460 -- Easier management and deployment of the various 461 VMs from the student’s point of view. 462 To estimate the grade of satisfaction from the students 463 with VMD, we conducted a survey on a random sample of 464 135 students. They were asked 3 questions: 465 -- Question 1 (Q1): Have you used the VMD software 466 tool to play the VM of the subject in the CC? 467 Fig. 4. Process of deploying a new MV -- Question 2 (Q2): Do you think it is easy to use the 468 VMD? 469 -- Question 3 (Q3): Has VMD allowed you to start 470 working faster than doing it manually using the user 471 interface of the virtualization software (for example, 472 VMware Player)? 473 474 The results of the surveys are displayed in Fig 5. As the 475 chart shows, the students consider that VMD has permitted 476 them to use their VMs in the CC quicker and easier than 477 using other methods, confirming the high degree of 478 acceptation of this tool by the students. 479 Regarding the satisfaction of CC staff and teachers, we 480 asked all of them. 481 The first group (3 members of maintenance team) was 482 asked whether the use of VMD solved the administration 483 problems caused by the use of VMs in the CC, fulfilling the 484 security requirements. We got a positive answer from all of 485 them. 486 As for the 10 teachers involved, the following questions 487 were asked: 488 -- Question 1: Has the use of VMD resolved the 489 problems of delay at the beginning of class sessions? 490 -- Question 2: What is your level of satisfaction with 491 the functionality of VMD (0 to 10)? 492 493 The results obtained were a unanimous “Yes” to question 494 1 and an average of 9.5 points for question 2, reflecting an 495 excellent level of satisfaction with the software tool. 496 VIII. RESTRICTIONS AND FUTURE WORK 497 The objectives for what the VMD model was developed 498 have been completely covered. VMD in its actual state is 499 totally operative, permitting the creation of portable 500 computer laboratories in CC with a simple user interface. 501 We aim to improve the software tool with a greater level of 502 universalization so it can be adapted to other possible 503 environments. The new version under development will: 504 -- Allow the student to save changes made in his VM 505 during each work session under two possible modes: 506 complete (creating a new image of the VM in the student 507 device) or partial (saving only the modified files via 508 “snapshots”). 509 -- Distribute and play VMs with distinct compression 510 formats (7zip, gzip, tar, etc.) or without compression. 511 -- Have the capacity to either copy the VM to the hard 512 drive of the computer or run it directly from a USB 513 removable media (no DVD-R). 514 -- Improve the user interface: although the current 515 VMD interface is user friendly, it is based on the 516 command line. Alternatively, a graphic user interface will 517 be offered for those users that prefer this working mode. 518 -- Play various VMs from a single invocation of the 519 VMD. 520 -- Support other virtualization environments in addition 521 to VMware Player, such as Oracle VirtualBox [14], 522 Microsoft Hyper-V [15] or QEMU [16]. 523 524 525 526 527 TABLE I 528 List of subjects that have used VMD 529 Subject Course Fundamentals and Applications of Telematics Services 2 Software Engineering 3 Advanced Telematics Services 4 Distributed Systems and Web Services 4 Databases 4 Design of Mobile Applications 4 530 531 532 IX. CONCLUSIONS 533 Virtualization techniques offer the possibility to create 534 flexible, versatile and low cost computer laboratories for the 535 teaching of practical engineering subjects. In this article we 536 have presented the VMD model that allows the efficient use 537 of VMs adapted to different subjects on a shared computer 538 facility. 539 From the educational point of view, this model has solved 540 the main inconveniences of teaching using VMs. VMD 541 simplifies both the management and deployment of the VMs 542 for the student, automating the whole process through a 543 simplified interface. In this way, it has eliminated the 544 reluctance of some students to use VM, due to problems 545 encountered with managing a VM per subject and with the 546 native interfaces used to start the VMs. At the same time, it 547 has reduced the delays detected at the start of class sessions 548 and therefore the academic performance of class sessions. 549 From a technical view point, it has given a solution to the 550 problem presented by the VM model when using a 551 particularized VM per subject in a CC. VMD allows the 552 adaptation of the characteristics of each VM to the hardware 553 particularities of the computer where it is wished to be 554 deployed and thereby optimizing its operation. Furthermore, 555 it adapts the VM to comply to the operational and safety 556 requirements demanded by the CC policies. All these 557 operations are done automatically and transparently to the 558 user, simplifying the tasks of professors (it is only necessary 559 to prepare a single VM model for the whole CC), CC 560 administrators (they don’t need to install the VM) and 561 Fig. 5. Satisfaction level of students with VMD students (they can use the same VM on any computer 562 regardless of its hardware characteristics). 563 The VMD has been used in several subjects in the last 564 few years with excellent technical results and student 565 satisfaction. 566 ACKNOWLEDGMENTS 567 The authors would like to thank the professors of the 568 Department of Telematic Engineering at the University of 569 Seville, José M. Fornés (Mobile Application Design), Isabel 570 Román (Software Engineering) and Juan A. Ternero (Data 571 Bases) for their constructive comments after using this 572 model in their respective subjects. This assistance has been 573 helpful to improve this software tool. 574 This work has been partially supported by the University 575 of Seville within the 2nd Own Plan of Teaching/Innovation 576 and Educational Improvement. 577 578 REFERENCES 579 [1] University of Seville, http://www.us.es/. January 2016. 580 [2] Bachelor of Telecommunication Technology Engineering. Higher 581 Technical School of Engineering. http://www.us.es/ 582 estudios/grados/plan_199?p=4. January 2016. 583 [3] "Bologna beyond 2010 - Report on the development of the European 584 Higher Education Area, Background Paper for the Bologna Follow-up 585 Group", Benelux Bologna Secretariat, Leuven/Louvain-la-Neuve 586 Ministerial Conference, 28-29 April 2009. 587 [4] Sierra, A.J.; Ariza, T.; Fernandez, F.J., "Establishment EHEA for 588 telecommunication technologies engineering degree", Global 589 Engineering Education Conference (EDUCON), 2012 IEEE Digital 590 Object Identifier: 10.1109/EDUCON.2012. 6201158 Publication 591 Year: 2012, Page(s): 1-6. 592 [5] IBM Tivoli Provisioning Manager for OS Deployment (Tivoli 593 Rembo), https://www.ibm.com/developerworks/community/wikis/ 594 home?lang=en#!/wiki/Tivoli%20Documentation%20Central/page/Tiv 595 oli%20Provisioning%20Manager%20for%20OS%20Deployment. 596 January 2016. 597 [6] L. Tao, L. C. Chen, and C. Lin, "Work in Progress - Improving Web 598 Security Education with Virtual Labs and Shared Course Modules", 599 40th ASEE/IEEE Frontiers in Education Conference, Washington, 600 DC. October 27 - 30, 2010. 601 [7] S. Ros, A. Robles-Gómez, R. Hernández, Agustín C. Caminero, and 602 R. Pastor, "Using Virtualization and Automatic Evaluation: Adapting 603 Network Services Management Courses to the EHEA", IEEE 604 Transactions on Education, Vol. 55, No. 2, May 2012. 605 [8] J. Hu, D. Cordel and C. Meinel, "Virtual Machine Management for 606 Tele-Lab IT-Security Server", Proceedings of the 10th IEEE 607 Symposium on Computers and Communications (ISCC 2005), 2005. 608 [9] P. Li, T. Mohammed, "Integration of virtualization technology into 609 network security laboratory", 38th ASEE/IEEE Frontiers in Education 610 Conference, Saratoga Springs, NY, October 22 - 25, 2008. 611 [10] K. C. Bandi, A. K. Nori, V. Choppella, S. Kode, "A Virtual 612 Laboratory for Teaching Linux on the Web", IEEE International 613 Conference on Technology for Education, 2011. 614 [11] VMWare virtualization tools, http://www.vmware.com/.January 2016. 615 [12] F. J. Fernández, A. J. Sierra, T. Ariza, G. Madinabeitia, J. M. 616 Vozmediano, Common Virtualized Environment in Telematic Labs, 617 International Journal of Online Engineering (iJOE), Vol 9, 2013, pp. 618 25-31. 619 [13] IEEE Std 1003.1™, 2004 Edition "Standard for Information 620 Technology - Portable Operating System Interface (POSIX). Shell 621 and Utilities", 2004. 622 [14] VirtualBox, https://www.virtualbox.org/.January 2016. 623 [15] Hyper-V, https://technet.microsoft.com/es-es/virtualization/.January 624 2016. 625 [16] QEMU, http://qemu.org/.January 2016. 626 627 Javier Muñoz-Calle holds a PhD in Telecommunications Engineering and has been a Lecturer at the Department of Telematic Engineering at the University of Seville since 2003. He has participated in many projects related to education at university level, including the implementation of PBL in engineering subjects, the development of ICT tools to support teaching or the use of game based teaching methodologies. His research interests include Security in Information Technology, Telematic Services, Cloud Computing, Advanced Management of Operating Systems and Web Application Development. Francisco J. Fernández-Jiménez has been a Lecturer at the Department of Telematic Engineering at the University of Seville since 2003. He has participated in many projects related to education at university level, including the implementation of PBL in engineering subjects, the development of ICT tools to support teaching. His research interests include ubiquitous processing, embedded systems, intelligent transport systems and mobile services. Teresa Ariza obtained her PhD in computer science from the University of Seville in 2000. She has been an Associate Professor at the Department of Telematic Engineering at the University of Seville since 1991. She has participated in many projects related to education at university level, including the implementation of PBL in engineering subjects, the development of ICT tools to support teaching. Her research interests include ubiquitous processing, embedded systems, intelligent transport systems and mobile services. Antonio J. Sierra has been a Lecturer at the Department of Telematic Engineering at the University of Seville since 1995. He has directed innovation projects in education, and has been responsible for partnership contracts with several companies in the ICT sector. His research interests include the IoT and computer languages, and the application of teaching methods. Juan M. Vozmediano has been an Associate Professor at the Department of Telematic Engineering at the University of Seville since 1993. He has participated in many projects of teleeducation at the university level, in the development of open educational resources and is a regular user of ICT tools to support teaching, having participated in the development of some of them. Outside the educational areas, his areas of interest include IoT, Smart City, Smart Energy Management and Systems.