Why On-Premises Solutions Are a Bad Idea
When it comes to getting product lifecycle management (PLM) systems up and working, Solution Consultant Mark Andres has managed both on-premise 'big box' PLM as well as cloud PLM implementations. Having seen PLM from both sides of the fence, Andres — without prejudice or harsh interrogation methods — confesses willingly to preferring cloud-based PLM systems. “The benefits of cloud are overwhelming, especially from an IT support perspective,” says Andres.
According to Andres, the benefits of cloud solutions can be divided into four key groups: installations, upgrades, patches, and maintenance. In this three-part blog post series, we will focus first on cloud solution installation benefits. Then — in what is sure to be one of the most anticipated PLM blog post cliffhangers since the 1980s when America wondered, “Who shot J.R.?” — we will examine how cloud infrastructure completely removes the need for companies to worry about upgrades, patches and maintenance.
Arena: Any complete enterprise system often consists of many moving parts to make it complete. There’s always a database layer, middleware layer, application layer, etc., along with security considerations. This would seem to provide some serious complications to implementing an on-premise solution, not for the faint of heart, correct?
Andres: Yes, plus you don’t just have one of these environments, you truly need at least two: a test/development environment and a production live environment.
This puts an added strain on an IT organization to accommodate all these technologies. Most companies do not have the depth of IT expertise in each of these areas.
Arena: Who are the people in the chain of command whom you had to work with to install and support a ‘big box’ PLM system?
Andres: First you have to assume these systems run on servers, usually multiple and often virtual servers. When implementing a client-based PLM system, I needed a systems administrator to build the servers. Typically, I would need three servers at a minimum: a database server, middleware/application server and a vault server. In one instance, the vault server needed to accommodate ~4 Terabytes of data, which at the time had to be broken into 2 Terabyte size logical unit numbers (LUNs). It is an expensive solution for many manufacturers.
Next I needed to install the database, which on its own needed a virtual machine (VM) server. I’d schedule my database administrator for that. And as often is the case, the database administrator wasn’t available for another three weeks because this person is tasked with many other projects, delaying this PLM implementation project.
Now you have to install the application layer. Find another IT person who ‘sort of’ specializes in these areas and get on their schedule. This also consisted of many pieces of IT equipment, the middleware component, then the application layer, any operating system (OS) needs like Java and don’t forget any ‘known’ patches to date.
Arena: Now, if you want a redundant system, which is typically spread across additional servers, that adds complexity, correct?
Andres: Yes. For instance, with this implementation, the vault software could accommodate 4 Terabytes of content, which wasn’t available at the time so we had to break this into two Terabye LUNs — again more complexity.
Arena: With on-premise solutions implementation questions are never ending. How much memory and disk space? How many drives? As well as dealing with database version, partitioning, archive mode, indexing, correct?
Andres: Absolutely. Lots of questions that have to be figured out before you can even get started implementing PLM.
Arena: What are some other little things that companies often overlook when implementing an on-premise PLM solution that can become painful to manage?
Andres: Oh lots of things like operating system (OS) support. Companies have to ask themselves many difficult questions:
- Can we support the OS on which the software runs? Do we have expertise in that area?
- Does the software require any Java components (from a Server and a Client perspective — both are important) and is it a version we can accommodate without impacting other enterprise applications?
- What browsers are supported with the PLM vendor? Since some ‘big box’ PLM systems do not support certain versions of FireFox, Chrome, Internet Explorer etc. it’s important to understand if your corporate IT structure accommodates the supported browser.
According to Andres, there are many cons to client-side PLM solutions. With cloud-based PLM, once your purchase contract is complete, you can have a PLM system up and running and configured for your business in as little as hours depending upon your complexity.
“I’ve been in cases where it is faster to order a new car and have it delivered to the dealer then getting some of these ‘big box’ PLM systems up, running and functional,” says Andres — an avid car enthusiast who owns some sweet rides in his garage. “And it goes without question that opportunity costs are less with a cloud-based PLM solution.”
To learn more about how cloud PLM removes opportunity costs associated with maintaining on-premise solutions, read this white paper “Why Executives Are Using Cloud PLM to Save Money”.