Not logged in » Login
May 30 2016

Folklore Debunked: 5 Server Myths That Are Hard to Eliminate


Despite serious warnings from their IT departments, many C-level executives and non-techies still seem to hold on to some popular beliefs about servers and data center infrastructures that may have rung true in the 1990s, but would provoke major outages today. We've picked up on the five most popular of these tales for a reality check and provide ammunition for the next time you have to enter negotiations with customers who want to run 24/7 services without investing in proper hardware.

Myth #1: "All Motherboards Are the Same"
Let's start with a classic argument that does in fact appear convincing at first glance. After all, pretty much every motherboard looks the same with its array of CPU socket(s), memory and expansion card slots, chipset, Flash ROM, disk and power connectors, controllers and all the other stuff that's attached to it, does it not? The problem here is that, just like in other walks of life, looks don't count for much – in other words, even though desktop and server motherboards may appear identical on the surface, they show very dissimilar characteristics when you look under the hood. Probably the most obvious and important distinction between these two classes of components is in their life expectancy: modern server mainboards are designed (and built) to last for five years and more on end in 24/7 operation, whereas standard desktop variants are typically built to run 10 hours per day and 5 days a week. So-called low-level components simply don't make it into our servers.

Myth #2: "A PC Will Do as a Server"
Until about eight or ten years ago, the world seemed to be full of hardcore hobbyists who couldn't help but relate success stories about how they had converted Mum's retired desktop into a bona fide server. While these accounts often seemed exaggerated (and somewhat repetitive), they did convey some kind of truth: back then, ambitious amateurs could easily build file, print or media servers using not-too-outdated PCs and the latest copy of their favorite Linux distro or Windows Home Server. However, as the latter name implies these configurations were only suitable for home environments or SOHO scenarios at best. Then again, lots of freelancers – e.g. lawyers, photographers or tax advisors – would switch to dedicated NAS filers and/or media servers as soon as they became affordable. And needless to say, they did that for a reason, namely because there's a clear distinction between adequate server hardware and components for PCs, as we pointed out above. Moreover, various applications such as mail servers, business intelligence/databases or OLTP require capabilities and performance levels that cannot be achieved with 'simple' CPUs designed for desktops and notebooks. Finally, genuine servers such as Fujitsu's PRIMERGY models can be adjusted to adapt to special needs, e.g. to enable individual MAC-addresses, virtualization etc. Regular, comprehensive tests ensure reliability and functionality way beyond what Mum's retired desktop could provide.

Myth #3: "Never Change a Running System"
It's a common belief that popular myths and legends are typically based on some kind of truth, i.e. evident historical facts or at least faithful reports that offer plausible explanations for past events. By these standards, the third assertion on our list is neither a myth nor a legend – but rather a fabrication that can be easily debunked. That's because the same people who recommend to never again touch a server once it works would never skip a security fix or system update for their own smartphone. And rightfully so, given the lamentable state of system, network and software security. – Servers for their part make up the very backbone of a company, since they carry out all business- and mission-critical tasks and then some. Simply imagine a bank or brokerage firm that's unable to execute financial transactions, construction or engineering companies that can't exchange data with their contractors and suppliers, or retailers and civil services where it takes ages to place an order or submit a request because server performance isn't up to par or because a system was compromised and must go offline instead of being revived within minutes. Let that idea sink in, and then proceed to imagine an outage or security breach occurred because the system in question was 'too small' or poorly maintained and patched. It certainly isn't hard to anticipate the anger and resentment of customers and clients who learn that they fell victim to a malfunction or had confidential data exposed because somebody wanted to save maintenance costs. Depending on the severity of an incident, this may even lead to a company or service being shut down. In light of this, adequate system maintenance including regular patching of a server's BIOS, firmware, OS and application(s) is not a necessary evil, but helps prevent unacceptable risks and lays the groundwork for running a successful business or service.

Myth #4: "Hot-Pluggable, Redundant Components Are Mere Cost Drivers"
Here we finally have an argument that is at least partially true. After all, an organization will indeed pay more for server hardware that's equipped with redundant and/or hot-pluggable fans, PSUs and HDDs. But make no mistake: system failures, data losses and service outages that occur due to a lack of such components usually cost a lot more money than the 250 dollars or euros a company may have to invest in a reserve PSU upfront.

Myth #5: "DIY Systems Still Offer the Best Price/Performance Ratio"
This last assumption also has a grain of truth in it, provided the CIOs/IT departments/administrators tasked with building these systems know where to buy adequate hardware components at a reasonable price and have enough skills as well as spare time to design and assemble all the bits and pieces. And even if they succeed, this says little if anything about the robustness and stability of a DIY system. In addition, many 'hand-carved' servers may theoretically have the CPU and RAM power to accommodate certain applications – and still fail to do so due to mishaps in design or configuration. By contrast, Fujitsu's PRIMERGY servers are regularly and intensely tested in collaboration with leading software vendors such as SAP, Microsoft or Oracle and certified for their applications. In other words: unlike the DIY model a PRIMERGY server offers reliable and stable performance from Day 1 and only requires minimal manual interaction because all required management and maintenance tools are included with the ServerView® Suite that's part of each PRIMERGY package.

Conclusion: Organizations that plan to build IT infrastructures for 24/7 operation should never rely on server myths first told during the 1990s that were often untrue to begin with. Following such outdated beliefs will only cause more troubles and costs than they can hope to avoid.

Marcel Schuster


About the Author:

Marcel Schuster

Senior Specialist Marketing Manager – Data Center Systems / Server


Comments on this article

No comments yet.

Please Login to leave a comment.


Please login

Please log in with your Fujitsu Partner Account.


» Forgot password

Register now

If you do not have a Fujitsu Partner Account, please register for a new account.

» Register now