New software will not run on existing computers.CHAPTER OUTLINE
The case of the unexpected upgrade
The case of migrating a legacy hardware platform to App-V or Hyper-V
The case of supporting a large application in a widespread organization
The case of managing a growing Data Center
Unprepared, Inc. has 300 manufacturing locations scattered across the United States. Their business is making widgets that they then distribute regionally from each plant. Each plant runs 24 × 7 and employs around 70 people (none of which are IT personnel). Supporting those workers are 20 computers deployed at each location and shared across three shifts. The primary application in use in the field is Widget Builder Pro v1.0 or WBPv1 for short. This program plays an integral part in their manufacturing process and they can afford very little downtime or they will miss shipping commitments and loose business to their competitors.
Unprepared, Inc.’s software vendor has announced that support for WBPv1 will end in 90 days and all customers need to upgrade to WBPv2. If customers do not upgrade they will no longer be supported when they have issues and new patches for WBPv1 will not be provided. To make matters worse the hardware requirements of the new version significantly exceeds the capacity of the 3-year-old Windows XP SP3 computers in place.
Management has decided that the cost and level of effort to deploy 600 new computers to run WBPv2 is not feasible and that running in an unsupported state is not acceptable. They now turn to the IT department to make things right and keep the business running.
New software will not run on existing computers.
Upgrading 600 computers is not feasible.
Running unsupported is not an option.
The conclusion that the IT department of Unprepared, Inc. made.
Terminal Services to the rescue!
After consulting with the vendor to ensure that they would support running WBPv2 under Terminal Services, the IT department went to work building their plan. Due to the number of users that would be running the program and the criticality of its availability, it was obvious that they would need a farm of servers. But how many servers would be needed to handle the load. The software vendor gave them some help by providing a general server specification and a number of users each server would support. Thinking it best to start small, validate, and then expand they put together their recommendation. They would start their TS farm with three servers and 10 pilot locations. They would build and deploy an .msi package that would seize the file association for the .wbp file from the locally installed WBPv1 to the remote-based WBPv2. They would deploy to each of the pilot locations one at a time taking both quantitative performance figures from their monitoring tools and subjective user experience interviews. Weighing these statistics and opinions along with the vendor guidelines they laid out an expansion plan so that management would understand when more hardware would be needed and what it would cost.
Well the rollout was a success, the downtime was minimal, they were never out of support with the software vendor, and the IT department at Unprepared, Inc. was heralded as heroes. OK, we all know that even the best plan still requires a lot of hard work to achieve its goals and that IT is rarely held up as heroes. However, it is part of the job that we all signed up for.
BigCompany, Inc. has a very old hardware platform supporting an outdated software application that has become critical to business functionality. Past attempts to upgrade the hardware or software have failed for various reasons; inability to get the proper amount of downtime, fear of breaking the application being hosted and being unable to get proper vendor support for the application.
There are two approaches that can be considered in this situation. In option one BigCompany utilizes App-V to create a virtualized instance of the outdated application. Doing so allows the engineers to take the time to properly install, configure, and validate the functionality of the application without having to interfere in the day-to-day operations occurring on the outdated hardware platform actively running in the production environment. Once the legacy application is validated in its newly “virtualized” form, it can easily be published to a current hardware platform running any flavor of Microsoft operating system. This approach allows for both hardware and software redundancy in the event of an emergency.
The second approach would be for BigCompany to fully virtualize the legacy server, complete with the software application already installed and configured. Doing so would also provide both hardware and software redundancy, with the only limitation (as compared to the App-V approach) being the ability to move the software application between various servers if needed.
BigCompany, Inc. has recently decided to make the version of Office Suite used by their employees to be consistent across the enterprise. However, due to networking and other infrastructure limitations, the traditional SMS delivery method is considered NOT to be an option for delivering the current Office Suite to all of the workstations that require it.
BigCompany has already begun designing their Dynamic Data Center to include application virtualization. The implementation of App-V allows BigCompany to package the latest Office Suite in a way that allows them to publish the application across the enterprise utilizing BITS and proper package creation. They include in the package the functionality to migrate the data and then uninstall previous versions of the Office Suite as part of the deployment of the App-V package to the workstation. Not only does this approach allow for the full deployment of the Office Suite to their entire enterprise, but it also allows for much easier updating of the Office Suite going forward by only having to update a single source and allowing it to replicate out to all of the workstations in a BITS controlled and silent fashion.
GrowingUp, Inc. has been increasingly successful since its formation not so long ago. The CIO of GrowingUp realizes that his data center will continue to grow as business continues to be successful. He also understands how quickly data center growth can make the management of critical systems and their data a losing battle for his engineers and their managers. The CIO makes the decision to plan ahead for future growth and management.
Creating a task force of engineers and technical leads, the CIO is able to properly design and implement a fully dynamic data center that will allow not only for proper growth, but also the proper monitoring of that growth and the systems and data involved. The results are a truly dynamic data center that is easily auditable and reportable, not only to the engineers who designed it, but also internal and external audit and regulatory organizations.
Undersized, Inc. was in a pickle. It needed to implement new accounting and payroll systems, but had nearly reached the capacity of their data center and could only add a few more servers. The two new applications would require 18 servers be added in order to support production, development, and testing environments. Undersized, Inc. had plenty of floor space open, but no employees had thought about the company’s power and cooling resources until they were faced with the grim prospect of pulling additional power and adding more air-handling units. The costs promised to be enormous and were not forecast into the company’s annual budget.
The IT department at Undersized, Inc. needed to devise a plan quickly that would support the new applications, not exceed the capacity of its data center, and be far less expensive than upgrading its power and cooling infrastructure.
The accounting and payroll applications were very similar in their hardware requirements. Each would require a database server and two application servers per instance. Two applications, times three instances (production, development, and testing), times three servers left them with a need for 18 new servers, but capacity for only six. Faced with the challenge of getting more out of less the IT team turned to thoughts of virtualization with Hyper-V. They worked closely with the software vendors to define their recommended requirements for memory and processors. They also confirmed that the vendors would support running their applications in a virtual environment.
The plan that they created was to reduce their new server count down to five machines. They decided to implement two production database servers, one for each application. This was based on performance concerns expressed by the software vendors with regards to virtualizing these machines. However, everyone was in agreement that the development and testing database servers could be run as Hyper-V virtual machines. They further concluded that all of the applications servers, production, development, and testing could be virtualized with Hyper-V as well. Basing their infrastructure needs on the vendors supplied hardware recommendations and with plans to thoroughly validate in their own environment, their design called for deploying three Hyper-V servers to run all of the virtual machines.
Ultimately, Undersized, Inc. was able to deploy the new applications within the confines of its data center by leveraging virtualization with Hyper-V to reduce the needed hardware footprint.