Decision-Making Metrics for Cloud-Based CAD12 Jul, 2017 By: Robert Green
CAD Manager Column: Determining whether cloud-based resources are right for your company can be confusing. For greater clarity, start your evaluation with this checklist.
When you want to buy a physical object, such as a car or a house or even a personal computer, how do you decide which one you’d like to buy? Chances are, you give the car a test drive, go walk through the house, or visit your local reseller and use the computer to see how you like it. Well when we want to purchase a cloud-based resource — say, a server, virtual workstation, or software product — it’s somewhat harder to make purchasing decisions, because you can’t lay hands on the device you want to purchase. So how can we all shop smart for cloud-based resources?
As Summer on the Cloud 2017 continues, I’d like to address the process of making decisions about which cloud-based resources could be right for your company — if any. In this column, I’ll share my strategies for evaluating cloud-based resources using fundamental business metrics and testing. Here goes.
A Few Examples and Conclusions
As with most complex problems, there are lots of questions to be asked and answered before a solution becomes clear. I’ve always found it useful to break complex problems into smaller pieces, then tackle each piece using common-sense metrics to draw conclusions.
To illustrate my method, here are a few simple questions I’m often asked that have complex answers:
- Should we consider using cloud-based workstations?
- How do we know if cloud-based software will work for us?
- Will data security become a problem if we use cloud-based storage?
- What are the intangibles in implementing cloud technologies that we should consider?
Of course, these seemingly simple questions are anything but simple to answer. I can offer some contextual conclusions, however, that may help you better understand how I approach solving complex problems:
There are no “one size fits all” answers. Cloud-based workstations may or may not work for you depending on your network topologies, distances from branch offices, and data center challenges, among other factors.
Testing and verification are more important than ever. You simply don’t really know how a cloud workstation, storage solution, or hosted software product will work until you try it. Even if something isn’t located in your office, you still have to test it.
Even though you can’t get your hands on them, you still have to take cloud-based solutions for a test drive before you commit. Otherwise, you may end up with a bad fit.
Each question brings on more questions. If you start with what you think is a single question — Will cloud storage be a security risk in your environment? — you’ll quickly realize that you’re actually facing multiple interconnected questions: What do your management team and client base think about security? Do you have contracts with health care or government firms that preclude cloud storage? What does your company’s business insurance stipulate about data losses? In this example, you’ll need to understand everything that contributes to your data security situation before you can determine whether cloud-based storage poses an unacceptable security risk for you. You’ll find that other cloud questions work this way too.
Network bandwidth and security are the great intangibles. While any manner of things could cause problems with cloud technologies, it is most likely network bandwidth limitations or security concerns that will be the showstoppers. Oddly enough, providers that base their solutions on a robust Internet connection never seem to consider that much of the planet lacks one.
You’d better understand IT. The more involved you become with cloud technologies, the more you must know about IT and IT policy. Like it or not, the cloud is an IT-driven environment, and for you to function properly you need to have solid IT chops.
Metrics Are the Answer
Given the conclusions drawn above, it’s clear that evaluating cloud resources is a difficult process. I’ve found that complex problems can be broken down using a checklist of known metrics to support smart decision making. While this prioritized list won’t answer everything in all cases, I’ve found it to be quite useful for my clients:
Client/contract requirements trump everything else. If a healthcare client or federal agency is required by law to keep all data on their own servers (and nothing on the cloud), then you can’t use cloud tools working on their projects — period.
Functionality comes next. Does the solution work optimally? Will the solution be workable given your project workflows? Will users embrace the solution? After all, if a solution won’t work for your needs or your users hate it, you’ll never be able to fully implement it anyway!
Does it save money? If not, why are you even considering it? If a cloud-based rendering tool keeps you from having to buy a $15,000 workstation, maybe it makes financial sense for you to use the cloud tool.
Does it save time (remembering that time equals money)? If that same cloud-based rendering tool could render all your project visualizations in 4 hours rather than 3 days and keep you from buying a $15,000 workstation, then it is well worth considering!
Does it reduce your support burden? If a cloud-based resource reduces your IT support, data center, and manpower requirements, then it could be saving you more time and money than you thought. On the other hand, if you must spend countless hours configuring, tweaking, and debugging the resource, it may be costing you money rather than saving it.
Could you work around resource failure? What happens if the Internet goes out and the cloud goes away? How would you work around this problem, and how long could your company tolerate the disruption? What would this cost?
If you think about it, these metrics are similar to those you’d use to consider buying non-cloud resources — and that’s exactly the point. Everything comes down to function and time (money) savings, as it should. The fact that the word “cloud” is in the discussion changes nothing when it comes to these factors.
Test, Test, and Test Some More
While many cloud-based technologies have great marketing pitches, you can’t afford to accept their claims without verification. Simply put, there’s no substitute for testing the product in your real-world environment. And as you perform testing, be sure to compile information on the usage, savings, and support metrics outlined above. Here’s how I approach testing, along with some things to note as you do so:
Get the vendor to agree to a realistic test. If the vendor doesn’t want to show you their product in an actual usage scenario, you should be concerned.
See exactly what’s required for setup. Do not let the vendor present you a slick demo that glosses over specifics; be sure you see all the work that it’ll take to make the resource work in your environment, and bring in your IT staff if needed. It is crucial to understand whether you have the in-house knowledge it’ll take to make the resource work.
Subject the resource to some user testing. Do your users hate it or love it? They’re the ones who will have to use it, so doesn’t it make sense to have them try it out?
Compare workflows based on time. How long does a particular task in your current environment take? What about when you use the cloud resource instead? Comparing the two is the only way to understand potential time/money savings in concrete terms. If a cloud-based resource offers no advantage here, why mess with it?
Understand how the resource fails. Will there be a way to work around Internet failure using the resource? Can you retrieve your data somehow if working offline is temporarily required?
As you perform testing, be sure you understand each of these variables and take good notes so you can quantify any time savings to be gained or setup time that may be required for computations later.
There are a lot of cloud-based resources out there competing for your attention, and it isn’t easy to understand what does and doesn’t make sense for your situation. My hope is that the strategies and metrics I’ve outlined will help you make smart decisions about cloud resources for your company. Until next time.
What are your thoughts on the CAD cloud? Please email me your thoughts at firstname.lastname@example.org and look for your feedback in the next installment.