Monday, 18th October 2021

6 common pitfalls to avoid when adopting virtualisation or Cloud

By Michael Thompson, Product Marketing Manager, SolarWinds.

With all of the visibility that server virtualisation gets, you would think that everyone is already using it everywhere. But the reality is that many people are still in the process of adopting virtualisation either for the first time in their organisation or for a new group, department, or application within their company. A study last year by Kaspersky Lab indicated that 69 percent of U.S. companies are currently implementing or have already implemented server virtualisation1, the flip side being that 31 percent of companies have not virtualised.

Dell and Intel did another study that showed adoption even lower in Europe, with the UK and Benelux leading the way with 41 percent and 43 percent of companies using the technology versus a low of 24 percent of Swiss companies using server virtualisation2. A VMware study in 2011 indicated that within companies, mission-critical applications were virtualised at only 40-60 percent, and all applications were an average of 59 percent virtualised3.

This all adds up to a lot of organisations that have not yet adopted virtualisation. If you factor in those that are going further and adopting cloud-based computing infrastructure, the percentage really drops.
The good news is that we now have enough information to understand the key pitfalls that companies often fall into as they begin the journey to either virtualisation or cloud. In this article, we walk through six of the most common errors teams make in their move toward virtualisation or cloud.

Common Pitfall #1 – Siloed Thinking
It is almost impossible in IT not to specialise to some extent. Being an expert in virtualisation, servers, storage, networks, and applications requires fairly substantial learning curves. As a result, organisations naturally begin to align people, tools, and processes with each respective silo.

A transition to virtualisation or cloud can truly be a disruptive technology shift. If parts of the organisation make the transition but others remain siloed, it can be disastrous. Virtualisation and cloud completely alter the rate of change required for an IT organisation. Modern hypervisors have done a lot to incorporate elements of server, storage, and network management into the hypervisor (e.g. datastores manage many related storage tasks, virtual switches can manage networking tasks), However, physical resources are still needed behind the virtual infrastructure. If a system is running out of physical disk space, there is no way around that other than provisioning more physical disk.

Teams moving to virtualisation and cloud need to think holistically, and align the broader team with the goals of the virtual implementation. Providing visibility via tools that can cross boundaries is one key way to bring teams together. If both the virtualisation and the storage team are looking at the same dashboard or alert monitor, they can better understand trends and problems to work together as opposed to the finger-pointing that often happens when an unexpected problem brings the system down. Also, revising procedures up front to account for the reaction time required by virtualised systems is another key factor in making sure that the broader IT team is bought into the success of the virtual or cloud implementation.

Common Pitfall #2 – Virtualising the Wrong Applications
Virtualisation works best for well-known, standardised applications. There is a huge range of experience, best practices, management tools, and vendor support for common applications like Exchange or SharePoint. In addition, the operations carried out by these systems are standardised and repeatable, and therefore are pretty well suited for automation.

Starting your virtualisation journey with a legacy custom application that is constantly being changed or modified is just asking for trouble. Other high performance applications like transaction processing or high-end databases are also difficult places to start.

Common Pitfall #3 –Buying into Vendor Marketing
C-Level or executive IT managers are constantly bombarded with slick marketing strategies, long-term visions, and glossy PowerPoint decks. With management holding the checkbook, IT admins can get stuck with hard to use, complex solutions that don’t meet their needs and that require external services for every change or upgrade.

So what is the IT admin in the trenches supposed to do? Be proactive! Perform at least a simple competitive analysis, research vendor tools, download and install trial product versions, and talk to colleagues at other companies that have moved to virtualisation or cloud. Always insist on being hands on with the technology in your environment for your use cases. Most (good) executives understand that they need to count on their people with deep domain knowledge for the success of a project. As a result, an IT team armed with lots of data up front can help set the right RFP requirements and have a major impact on the ultimate technology selection.

Common Pitfall #4 – Not Setting a Baseline or Goals
Change is always difficult. As people adjust to new systems and approaches, there will always be those that long for the old ways. They often express this longing with statements like “it worked better with the old system” or “the new system is too slow.” The best way to fight this natural negative reaction to change is with data.

Before you transition to a virtualised system, be sure to identify key metrics and benchmarks that allow you to baseline the existing systems before you transition or decommission them to the cloud or virtualised systems. Then you will need to collect the analogous data for the new systems. This allows the data to do the talking and avoids arguments based only on subjective opinion. In the case where skeptics are actually right, you have data to identify and correct problems early before broader support for the initiative is jeopardised.

Once you have a baseline, set quantitative goals to track you own progress with the project. It can be tricky setting goals for a new system where it is hard to know exactly what the performance will be. One good approach is to stick to higher level business goals for the broader audience – time to deploy a new application, labor cost per server managed, etc. Internally, you may want to set more detailed performance goals to use as a guidepost to help identify problem areas to work on.

Common Pitfall #5 – Inventory Skills
The staff running the virtualisation or cloud environment can make or break a project. Going in, it is critical to have the skills you need to be successful. Based on a virtualisation skills inventory, have a plan to develop adequate skills coverage. This could include training classes, hiring in expertise, mentoring highly skilled with entry level employees, etc.

Supplement the skills by documenting the domain knowledge you learn through the process. Set up procedures for common operations and best practices to capture the “tribal knowledge” you generate as you pilot and implement virtualisation or cloud environment. This also makes you less susceptible to problems if your lone VMware expert takes another job or goes on vacation.

Common Pitfall #6 – Thinking Easy is Always Good
Easy is almost always good, but with virtualisation and cloud there can be a downside to easier and faster. When new VMs can be deployed in minutes or a new Amazon cloud server is purchased with a credit card, many of the natural limits of physical systems are gone. While “VM sprawl” is not a new concept, the time to deal with it is as you are setting up you environment—not after you have discovered that you have a problem.

If you set expectations and parameters up front by telling your users that they only get a certain number of VMs, a limited volume of storage, or that VMs that haven’t been logged into for three months will be deleted, it will be much less painful for you and your users when it comes time to start enforcing those limits. Imposing them after the fact means that users have to go back to VMs and files that are old and try to remember which are important, which to archive, and what gets thrown away.
There are enough other mistakes to make when starting down the path to virtualisation or cloud; there’s no sense making the most common ones.

Michael Thompson is a product marketing manager for SolarWinds’ application and server management product portfolio.

1 Survey: Among U.S. Companies Adopting IT Virtualization, 53 Percent Use Sub-Par Protection to Secure Virtual Infrastructure;

2 Benefits of server virtualization more significant than the obstacles;
3 VMware Journey Adoption Insights 2011 Report;

he EMEA external storage systems market value was up 3.3% year on year in dollars but down 5.5% in e...
The University of Aberdeen’s Biomedical Imaging Centre has successfully deployed BridgeHead’s Indepe...
Revamped channel program doubles down on Platinum solution providers, increasing incentives for new...
A Forrester Consulting study commissioned on behalf of Dell Technologies, including data decision-ma...
Spectra Logic has introduced several enhancements to its award-winning StorCycle® Storage Lifecycle...
Chris O’Brien Lifehouse becomes the first hospital in the world to implement BridgeHead’s RAPid™ Dat...
DataRobot doubles down on MLOps with remote model lifecycle management agents, automatic deployment...
NetApp has unveiled a new release of its ONTAP software that powers data services on premises and in...