Top 10 hardest Higher Ed software titles to deliver

Some software applications are simple to deliver, both technically and commercially. Others require more consideration, higher investment, and can be quite problematic for Higher Ed IT. This article will cover some of the hardest-to-deliver software applications alongside considerations to be made when delivering, from packaging to provisioning.

To deliver on-site to a specialist lab with sufficiently powered machines, application virtualization may be the way to go; this technology leverages, and is executed on, the hardware of the endpoint. This reduces the need to make VDI all the more expensive by having to invest in further GPU capabilities. Conversely, if your VDI estate already benefits from extensive GPU capabilities, utilizing VDI may help save on heavy investment into endpoint hardware, with the extra benefit of making the software more universally accessible across campus rather than exclusively in specialist labs. It is worth noting that virtual desktops still need to be imaged. Keeping images as small as possible and minimizing the laborious task of imaging is another reason to consider application virtualization.

Off-site delivery may require a little more provisioning logic, depending on the information you have about your end-users and their hardware. For engineering students who are used to using CAD/CAE and likely have the appropriate hardware to run such applications, application virtualization would be beneficial as the chosen software delivery method. It can leverage the hardware of their own device, doesn't require a constant network connection as VDI does, and IT benefits as application virtualization is an inherently cheaper delivery technology than VDI.

SPSS is a medium-weight software application with no GPU requirements. The result of this is that SPSS can be delivered via VDI without any further investment in GPU server hardware. However, VDI is unlikely to be the most cost-effective method of delivery and may not guarantee the best user-experience. VDI's reliance on network connection means performance is subject to network data speeds. Application virtualization makes use of end-point hardware; 1GB of RAM is not a big ask when it comes to contemporary hardware; the vast majority of modern laptops possess as much RAM, if not more than this. This makes application virtualization perfect for on-site delivery. While 800MB isn't a lot of disk space, applications of this size quickly start mounting up when included in images. If delivering via imaging, perhaps only include in golden/master images specific to specialist statistics/math PC labs.

For delivery off-site, all of the above applies. An exception to this that may mean VDI is more suitable is if your SPSS license agreement stipulates that it must only be executed on-site. VDI means that software is run on-site and then pixel-streamed to end-users, keeping delivery compliant with stipulations such as this one.

For delivery on-site, it is not advisable to include MATLAB in your images. With a minimum of 3GB required for MATLAB only, and up to 31GB for a full installation of all MathWorks products, using imaging to deliver this program will quickly result in image bloat and long, laborious image rebuilds should any changes need to be made. Low GPU requirements and high demand for disk space perhaps make VDI more appropriate for MATLAB delivery than for previous entries on this list. However, many of VDI's benefits lie in delivering to BYO devices and some may consider it a 'waste' of expensive VDI licenses to deliver software with VDI on-site. Application virtualization is once again advisable providing end-point hardware has enough disk space and is not a zero client or ultrathin client.

VDI may be required to deliver MATLAB off-site to non-managed devices if there's a chance that student/faculty-owned devices do not meet MATLAB's RAM and disk space requirements. Once again, if your license agreement stipulates that you must run MATLAB on-site only, then VDI should be a consideration. These things aside, application virtualization is the best objective choice of delivery technology for delivering MATLAB.

• Main challenge: A very large app with high computing requirements.
• Recommended delivery on-campus: To enable a universal lab approach across campus, delivery with AppsAnywhere and Cloudpaging is recommended.
• Recommended delivery for BYOD: For quick and easy deployment to student-owned devices, delivery with AppsAnywhere and Cloudpaging is recommended. For use on devices that don’t meet the high hardware requirements, seek an alternate delivery method below.
• Alternate delivery methods: For some use cases, campuses may opt to provision a high-end physical computing lab, VDI, or server-based delivery of ArcGIS, however, each alternate method has its own drawbacks.

Once again, for delivery on-site, it is not advisable to include an application of this size in an image or master image unless there is no other option. VDI may be necessary to deliver to zero/ultrathin clients as RAm requirements are fairly high for ArcGIS titles. However, aside from this, application virtualization is the best and most cost-effective method of delivering to on-site, managed hardware with the capabilities to handle it. Virtualized applications act exactly as if they're locally installed and so an application like ArcGIS which may be operating with many different plugins, integrated titles, and/or dependencies would benefit from this feature.

This is mirrored for off-site delivery to BYO devices, and VDI should be reserved for where the end hardware is lower in spec and therefore not able to run ArcGIS. However, the vast majority of student-owned devices are likely to be able to support the RAM and GPU requirements to run ArcGIS, and so using application virtualization will provide the best user-experience and most cost-effective solution to delivering ARCGIS.

• Main challenge: A very large suite of apps with high computing requirements.
• Recommended delivery on-campus: To enable a universal lab approach across campus, delivery with AppsAnywhere and Cloudpaging is recommended.
• Recommended delivery for BYOD: For quick and easy deployment to student-owned devices, delivery with AppsAnywhere and Cloudpaging is recommended. For use on devices that don’t meet the high hardware requirements, seek an alternate delivery method below.
• Alternate delivery methods: For some use cases, campuses may opt to provision a high-end physical computing lab, VDI, or server-based delivery of Autodesk, however, each alternate method has its own drawbacks.

If the software titles need to be imaged due to the capabilities of lab computers/clients or for VDI delivery, it may be worth considering keeping this separate from a general master/golden image. this might mean that specialist labs are necessary. If the image is already built and you're looking to reduce specialist labs, making a certain number of Autodesk instances accessible via VDI would help in this respect. For machines with the hardware to run these titles, application virtualization is preferable. This extends to delivery to BYOD devices, where capable devices should be delivered to using application virtualization and VDI should be used as a last resort.

Packaging and delivering such a huge number and variety of software titles with inter-dependencies and integrations with each other is a tough task; due to the applications covering such broad range of functions, it's unlikely they can easily and efficiently be arranged into specialist images without the number of images becoming inordinately high. It is because of this that application virtualization is the first and best solution for delivering the entire suite. Virtualized applications are executed in virtual structures separately from the end operating system and can interact with plugins and other programs seamlessly, exactly how they would do if everything was locally installed.

For the same reason as previous entries in this list, NVivo isn't a great candidate for imaging given its high demand for disk space. With no GPU requirements, NVivo could be delivered through VDI, however, its low demand for CPU performance makes it the perfect candidate delivery via application virtualization, whether the end-point is a managed device or a BYOD device. It will function well on all but zero clients, will be useable without an active network connection, and will provide the best user-experience possible for the end-user.

Moderate in size and performance requirements, the only real use case for VDI in delivering Maple is if the license agreement stipulates that it may not be delivered off-site or delivery to Chromebooks or non-compatible devices. You may want to include Maple in your master image for relevant departmental labs and learning areas, but otherwise, application virtualization is the strongest contender for efficient delivery with universal access and a strong user-experience.

Adobe is historically notorious for its license agreements making things difficult when it comes to delivering in Higher Education, particularly to off-campus devices. Using named user licenses/single-user licenses has been known to cause difficulty, as it prevents the traditional approach of pulling from a 'pool' of licenses as-and-when users request them, making on-demand and universal access a little more difficult. It does, however, help to make off-campus delivery potentially cheaper, as users may be able to login and download Adobe apps directly to their machine; in some cases, IT may only need to provide a link to Adobe CC download or specific Adobe apps. In cases where Adobe's licensing forbid delivery off-campus, VDI will need to be leveraged to run and execute the software titles in a server on-campus, and then pixel-stream the application to the end-device and requesting user.

Alongside browsers and PDF readers, Microsoft Office is one of the most popular and commonly used software suites in all of computing, let alone Higher Education. Despite it having many options for delivery and sometimes not necessarily needing a separate delivery technology, it is one of the most asked about software packages, in terms of how to deliver it, due to its universal necessity. For on-site delivery, Microsoft Office is one of the few recommendations to include in your imaging, as well as the aforementioned browsers and PDF readers. 

To deliver Microsoft Office off-site, there are a few options. Firstly, it is definitely worth avoiding using expensive VDI licenses to deliver Office given the sheer number of methods there are to get this to your students when they are off-campus. Office's AD integration allows access through SSO, meaning your students can download the individual titles directly to their machines and simply use their student email details to access. The web versions of these products can also be accessed in the same way. If these options aren't viable for your university, then application virtualization is the way to go. Office is non-demanding and small in file size; application virtualization will provide great user experience, doesn't require a constant network connection, and is an inexpensive method of delivering to off-campus devices.

A commonly asked question in software delivery is how to deliver plugins, or deliver applications in such a way that they'll function properly with any given plugin. Application virtualization is the answer. Using app virtualization, software is virtualized in a virtual structure and executed separately from the operating system. However, apps still look, feel, act, and function as if they're locally installed permitting seamless use of plugins, other integrated programs, and system context menus.