At the current time, WLab requires the following:
While the application requires access to a Hyper-V server (which inherently requires a Windows Server installation), WLab itself can be hosted on any platform, and has been tested on all three of the major operating systems (i.e. Windows, OS X and Linux).
The WLab application must be run on a server that is on the same physical network as the Hyper-V server, with no intermediate routers in between. Ideally, your configuration should have the Hyper-V server and the Tomcat server that runs WLab both allocated static IP addresses. VMs should then obtain IP addresses via DHCP on the same subnet. For example:
| Hyper-V server: | 192.168.1.5 | |
| Tomcat/WLab server: | 192.168.1.6 | |
| Jumpgate server: | 192.168.1.7 (this is explained later) | |
| DHCP range: | 192.168.1.10 – 192.168.1.210 (gives room for 200 VMs) |
Where possible it is recommended to allocate a dedicated Hyper-V server to WLab, simply because of the number of dynamically created virtual machines involved.
Consider a course with 20 students. The tutor has created 8 lab exercises for this course, each divided into 4 stages. This means that there will be a maximum of 8 x 4 = 32 tutor VMs. These will then be cloned for each student, i.e. 32 x 20 = 640. Thus the total number of VMs that will be created for this – rather small! – course is 672.
This number is not as enormous as it may sound in terms of system resources – assuming that the students only run 1 VM at a time, the worst case scenario is that 20 of VMs would run concurrently. The VMs also use differential disks, so the amount of disk space occupied will not be excessive. However, if this number of VMs were created on a production Hyper-V server that also contained other, non-WLab VMs, it would become impossible to administer these non-WLab VMs among the “noise”.