Where to find "best practice" for any given IT scenario
-
Official documentation for whatever you're trying to configure is a good place to start, then cross reference that with what smart-sounding people who have used it before say works best.
-
Surprising number of responses with how quickly I posted this.
Glad to see most people agree that Google is a top listed item, but the multiplicity of results is often far to much, or misleading.
Reading vendor documentation doesn't necessarily fill the "best practice" position, as it may only be "best practice with our product"
Posting topics on SW or even here for example, are great for discussion but doesn't always result in a great answer, or an open ended answer.
All of the above are good places to look, but aren't always clear cut.
@scottalanmiller SMB IT Journal does appear to be a great place to get a better understanding of a specific subject matter. I've read several articles there my self and it makes complete sense to me. But could it really contain the majority of Best practices in the IT and Business world?
-
You also have to watch out (especially on SW) for discussions started as "what are best practices for ____", when what they actually mean is "please walk me through literally every step of how to _______ , and please dumb it way the hell down for me". But that is where critical thinking skills come in...
-
That's just it, there is no best practice for anything but the most basic of tasks. It's almost always scenario driven.
That open ended situation you talk about on SpiceWorks, or here - that's what shows that there is no one best option/opinion.
-
@Dashrender said:
That's just it, there is no best practice for anything but the most basic of tasks. It's almost always scenario driven.
That open ended situation you talk about on SpiceWorks, or here - that's what shows that there is no one best option/opinion.
Although there are often best practices per scenario. These are called implementation patterns. Software Engineering really drove that concept hard with the GoF thesis.
-
spiceworks
-
@scottalanmiller said:
Although there are often best practices per scenario. These are called implementation patterns.
The trick, I find, is to figure out how the best practices actually fit into your scenario... No two setups are exactly the same, even if built on the same hardware platform(s). In a lot of cases, I think it may be possible to modify your scenario to fit industry accepted best practices. Whether or not that is wise / more costly... I'm not sure.
-
There's also a balance between following best practice and thinking outside the box. A lot of progress has been made as a result of mavericks who were creative and didn't follow accepted best practice. Indeed, best practice changes over time, often as the result of people trying new ideas. There will always be times when I like to go my own way rather than follow the crowd, even if that often turns out to be a mistake.
-
@Carnival-Boy said:
There's also a balance between following best practice and thinking outside the box. A lot of progress has been made as a result of mavericks who were creative and didn't follow accepted best practice. Indeed, best practice changes over time, often as the result of people trying new ideas. There will always be times when I like to go my own way rather than follow the crowd, even if that often turns out to be a mistake.
In that scenario, the best practices are the place to base things on though.
BP says X
You try Y and see how it compares. -
Google, but how about if you create your own best practice?
-
@Joy said:
Google, but how about if you create your own best practice?
Isn't that kind of the opposite of best practice? Now you can certainly make your own BP based on a specific scenario of your own, but then it's unique to you.
-
@Joy said:
Google, but how about if you create your own best practice?
You can't create best practice in just one scenario. The solution has to be tested and tried in multiple scenarios for it to become a "Best Practice".
Otherwise it's something that we've seen done once before, and the results of it are unverifiable.
-
@DustinB3403 said:
@Joy said:
Google, but how about if you create your own best practice?
You can't create best practice in just one scenario. The solution has to be tested and tried in multiple scenarios for it to become a "Best Practice".
Otherwise it's something that we've seen done once before, and the results of it are unverifiable.
But if you are brave... then you can do it again and see if you get the same results as the first time. 8-)
-
My issue with best practice is that it often requires spending a lot of money. Some of my more creative solutions have been motivated by saving money. Sometimes cheap solutions are a false economy, when they become so flaky they end up costing more to maintain and fix than just doing the job properly to start with, but sometimes they are successful.
But I'm also not entirely clear on the difference between "best practice" and "industry norm". For example, the industry norm is to run Microsoft Office, but that doesn't make it best practice. Best practice may be to run LibreOffice, but not many companies do. If you started a thread on Spiceworks saying which Office suite should I use, I reckon almost everyone will say Microsoft Office. Does that make it best practice?
-
@Carnival-Boy Using a specific application such as Microsoft Office vs LibreOffice for an example is a very poor example.
RAID 5 vs RAID 10 on spinning rust is a much better example.
RAID 5 was Industry Norm several years ago. Now RAID 10 is because it's much more resilient to URE's as well as other failures.
-
@DustinB3403 said:
@Carnival-Boy Using a specific application such as Microsoft Office vs LibreOffice for an example is a very poor example.
Sorry, it was the best I could do
I'm basically a software guy with little interest or knowledge in hardware issues like RAID. -
@Carnival-Boy OK think of it like this, the software comparison is like comparing two cars, both do the exact same thing, get you from point A to Point B.
Both are reliable enough to work consistently.
So there's really no comparison, at this point its simply preference as to which you want and what you have to spend.
Comparing the example to something that is measurable, lets say the cars crash test rating would be a better example.
Car A gets 2 starts for front end crashes. Car B gets 5 Stars.
Which car do you want to drive?
-
@DustinB3403 said:
@Carnival-Boy Using a specific application such as Microsoft Office vs LibreOffice for an example is a very poor example.
RAID 5 vs RAID 10 on spinning rust is a much better example.
RAID 5 was Industry Norm several years ago. Now RAID 10 is because it's much more resilient to URE's as well as other failures.
That's a bit different. RAID 10 isn't a "best practice", although avoiding RAID 5 is. The best practice is to know RAID and evaluate your needs on an individual bases.
What you are looking at here is a "rule of thumb" or a "de facto starting point." If you know NOTHING about storage and need storage for your server, start from OBR10 and diverge as needed. Best practice is not to know "nothing" about it and make informed decisions. OBR10 gives you a default, a place to assume "most common" but that's different than a best practice.
-
Best Practice: A routinely repeatable process to achieve best results. Should be applicable at least 99% of the time and probably far more.
Rule of Thumb: A common and general safe assumption as a reasonable way to go when you don't have time or the ability to thoroughly assess all of the factors. Should be applicable more than 50% of the time. And should err "on the safe side."
-
@DustinB3403 said:
RAID 5 was Industry Norm several years ago. Now RAID 10 is because it's much more resilient to URE's as well as other failures.
RAID 5 was a norm but never a best practice. Even in the 1990s it was presented as a reasonable risk to save money. Not the same as a best practice by any stretch.
