My boss asks: why FreeBSD?

Hi,

Please, I need some help. My boss only knows Linux and asks me "Why now choose FreeBSD?" With a new OS like FreeBSD, some people need to learn this new OS, it takes some time to do that.

So, how can I explain to my boss that FreeBSD is a good OS? How do I justify to teach to some people (1 or 2) how to use FreeBSD?

Speak about Jails? ZFS?


Thanks for your help
 
If I should pick one argument, I will use either great documentation (for example compare FreeBSD and Linux man page for ifconfig) or stability over time - in network with hundreds of Linux wireless routers I never knew, if I should use ifconfig, iw, wi, iwpriv, ip etc. Some interfaces created with an ip can not been seen with ifconfig utility as bonus etc. Also instalation of 3 different incompatible versions of fdisk during use of the official Debian embedded image creation script, which even does not work without rewriting half of that mess (can't even locate own config files) says something.
 
Where to start?

For me:
  • Clear filesystem organization
  • Clear configuration structure and approach
  • Focused on performance, robustness, and security and delivers on it
  • Not at the whim of "backyard programmers"
  • No GPL entanglements for the OS itself

Too many times in Linux that I've configured something using one GUI tool and gotten totally hozed trying to do something else (for example, network configuration past "use DHCP" or "assign a single fixed IPv4 address to the interface")

The clean separation between OS-level configuration and scripts (in /etc) and "add-on" configuration (in /usr/local/etc) is beyond value, especially when upgrading.
 
Or you can go straight to the point.

Linux%2Bvs%2BFreeBSD.jpeg
 
nORKy said:
"Why now choose FreeBSD?" With a new OS like FreeBSD, some people need to learn this new OS, it takes some time to do that."

So it does for the script kiddies and drive-by intruders.
We need maybe some days to adjust, if at all.

Someone trying to root a box with a buffer overflow script from some leet toolbox bumps his nose and then? Sitting down there and spending some time to adapt?
Attackers who would do that, there is no real protection as these also would lockpick your office and steal the hardware.

More or less quoted from an admin we had who had a gateway running some seriously obscure combination of PF on a (emulated) VAX.
 
I think it depends on what services the company is providing. For instance, if your company is selling things, you need to talk in sales-terms to your manager and convince him that FreeBSD is able to "sell more" by providing better stability and better performance that Linux on the same machines. From my experience, it will help nothing if you talk to a non-techie person in techie terms. Bottom line is that you need to talk in his language for him to understand you.

My 2c.
 
As @da1 pointed out: It's hard to say - depending on your environment. What are those Linux servers doing? In some cases, unfortunately, you can't use FreeBSD. SAP application server is one good example, oracle hosting server is another.
These are all technical questions and are more or less not important for management. Usually it's green/red cell in excel whether this solution is acceptable or not.

But there is also question about the support/SLA. Do those (linux) servers come with it? It's really bad if you hit a bug, system keeps panic-ing and you can't make heads or tails of it. Quick and prompt support (and patch for that matter) is a need for the business.

I've never seen such support in FreeBSD, but that's because I never looked (I guess some 3rd parties may do this).
 
da1 said:
"... if your company is selling things, you need to talk in sales-terms to your manager ... "
"From my experience, it will help nothing if you talk to a non-techie person in techie terms. Bottom line is that you need to talk in his language for him to understand you."

My 2c.

You are very correct in this, sadly. I for one find it really annoying when you need to powerpoint your information instead of simply telling the other side of the communication about your arguments. So you need to learn to speak "marketing", which might as well be it's own language. (Since marketing seems to live on it's own planet, that is understandable, isn't it? SCNR)

So what is there to do for you?
  1. find out what you need
  2. find out what you can have
  3. find out what you want
  4. list pro/contra points for each
  5. ... also as seen from the other side of the table
  6. follow the gradient

The point about listing pro/contra is extremly hard if you are not used to it. I found that several years of D&D and other role playing experience can help you to ignore what you know and see some situation from the point of someone else.
That is why I sometimes come overdressed to a meeting because I know that the customer to talk to has no technical background. So he tends to stick to things he trusts - and that may be a three-pice and cufflinks. For other customers, the t-shirt, sneakers and cargo pants are better. That are those who come from the same stable as you, who have a solid technical background and who also see a tie as a competence simulator.

Long story short - make your homework, the other side of the argument needs to trust you to take your information, and it needs to like you to act on that information.

But in the end it comes down to numbers. If you convince your boss of a solution that does not work out, guess who gets the heat for it. Be sure to move into a direction where you want to go, where it needs to go and where real benefits for all involved are possible.

My first answer addressed the uptime benefit and the better protection from random problems by script kiddies. These can be real benefits for a small biz, but they do not protect from dedicated evildoers.

What you (or others who have to deal with customers and sales/marketing) may want to do is spend a weekend with Sun Tzu, The Art Of War. If nothing else, it will help you to understand those guys better.
 
nORKy said:
So, how can I explain to my boss that FreeBSD is a good OS? How do I justify to teach to some people (1 or 2) how to use FreeBSD?

Because, they will have a great learning experience! And as a good boss he wants to have smart employees, wouldn't he ;). As a side effect, from learning FreeBSD and opening their minds they will probably also learn to understand and troubleshoot the Linux systems much better.
 
business reasons, not techie

Provide real reasons to your argument using the business language, is not so hard as you think.

Examples:
  1. We are going to reduce the time it takes to deploy websites because fbsd FreeBSD enables us to script / automate the process faster because .........
  2. We are going to increase the security of our servers due to design of the fbsd FreeBSD OS
  3. We are going to reduce the time it takes to deploy a new server because fbsd FreeBSD will let us ...... blabla
  4. We are going to recover from hardware failures because fbsd FreeBSD enables us to ....... blablalbla
  5. We can implement solutions that we don't have today because with linux it takes more time

Basically he wants to hear about cost reduction, increasing potential of existing hardware, implementing more solutions/services, etc... It will be your task to really make that difference with FBSD FreeBSD.
 
Has there really been any conclusive OS vs. OS comparison of Linux and FreeBSD done in the past 2-5 years? Obviously FreeBSD nor Linux can live on its merits from 10-20 years ago. I think a good project for FreeBSD advocacy would be to do this on current hardware under a range of enterprise and desktop workloads.

Actually this is something that is relevant to my interests, so if I can conjure the time/effort/Red Bull I may try to get something together.
 
shitson said:
Has there really been any conclusive OS vs. OS comparison of Linux & FreeBSD done in the past 2-5 years? Obviously FreeBSD nor Linux can live on it's merits from 10-20 years ago. I think a good project for FreeBSD advocacy would be to do this on current hardware under a range of Enterprise & Desktop workloads. Actually this is something that is relevant to my interests, so if i can conjure the time/effort/Red Bull i may try to get something together.

Last year I did some sysbench tests of FreeBSD 8.2-STABLE+ ZFS + aesni vs Debian GNU/Linux 6.0.2 Stable + xfs + aesni on i3-12xx ~ 2,4GHz / 8G RAM / 2x500GB WD RE4 / Supermicro MB machine. FreeBSD was about 4 times faster with 8 MB sized files in random read write. On the test with 80MB sized files the two setups were almost equal. On 400MB sized files Debian GNU/Linux was a lot faster.

Need to search for the test records for precise results.
 
jrm said:

In all due respect to objectivity you can't judge it on nearly 7 years old paper.

I did, however, find some other, similar, newer (2012) papers on developers @ IBM page, but I don't think it's enough. IBM is not offering it in their services as they offer Linux (very well known setup in production - HP-UX/AIX/Solaris with Linux application servers as their NFS satellites). Neither does HP.

It's true I never saw FreeBSD connected to a SAN dealing with multipathing, but I can confirm it's a nightmare in Linux (compared to HPUX/AIX e.g.).

Needless to say I would love to see FreeBSD used here instead of Linux. Maybe some day ..
 
My reasons for using FreeBSD really don't have much to do with performance. It performs well enough, I don't care if it's a few milliseconds faster in some benchmarks or slower in others. In the end this almost always comes down to system/software tuning to really get performance.

After getting shafted by Redhat 8/9x I swore I'd only go with community supported distributions. You can never be sure when a company decides that you're "no longer commercially viable" and drop you as they please. What I like most about FreeBSD is that it is stable and consistent. It has a well defined division between the system, and third party software. I've had problems with upgrades sure, but I've always been able to fix them. Sometimes from some pretty messed up states. MS Windows doesn't expose the guts enough to make that possible, nor does it provide you with the tools to do that kind of thing. Linux has the tools and also does expose the guts, but what about the documentation? Maybe it's my imagination but it's hard to find answers to Linux related questions, because many distros do things their own way, use different sets of cobbled together software to implement the system, or the documentation you find is severely out of date. I still can't get Gentoo to boot into single user mode, nor can I find the correct answer to get it to do so.

The more I've learned about FreeBSD, the more I've been able to leverage the system. Learning Windows always made me feel like learning the latest techniques to apply band-aids and put up with the latest "great MS initiative", while learning Linux often feels like I'm learning the latest in fashion, only to be dropped when the next shiny idea looks better.

FreeBSD isn't the best for everything. I wouldn't go that way for a desktop, but for a server it's always my preference.
 
There are many comparisons around: Linux vs BSD vs Windows. The problem is that they are, for the most part, terribly outdated. It drives me mad when people keep mocking Windows for its blue screens. I have to use it in the office - and I can't remember a BS for nearly a decade now. This is all long gone.

What I like about BSD, is that it is more UNIX than Linux, technically/historically it IS Unix. And it simply feels more stable, robust. I like simplicity. An OS should NOT put you into a GUI. This would be plain wrong. ;-) It should not hide everything under config tools.

Performance-wise it seems to be pretty hard to judge who's in the lead. After all it is emphemerical. Just a snapshot, I guess. Like Michael Lucas said it is a tug-of-war, with different annual outcomes..
 
My biggest attraction to FreeBSD is its consistency between versions. All the Linux Distros seem to have their own idea on where stuff should go, missing the point that it's better if everyone puts stuff in the same location than finding a better location, but being the only one to use it.

FreeBSD is also nice in terms of being a barebones OS, giving me the very standard tools and letting me decide if I really need the latest and greatest tool for that job (rsyslog vs. syslog) or (anacron vs cron) - One gets lost in determining what version is running on a base install.

The thing that really deters most from the OS is its seemingly large learning curve, being confronted with no X11 really scares some people off as well. But most miss the Documentation Project and the vast array of knowledge that you can gain from it.
 
shitson said:
My biggest attraction to FreeBSD is its consistency between versions. All the Linux Distros seem to have their own idea on where stuff should go, missing the point that it's better if everyone puts stuff in the same location than finding a better location, but being the only one to use it.

This.

My "hardware" platform these days is vSphere, so Linux driver support holds no advantage for me.

I'm sure I'm not the only one who's pretty much phasing out bare metal installs for most stuff because unless you're playing at the extreme high end of user numbers, modern hardware is way more than fast enough to support multiple VMs on a single machine.

You also get failover, snapshots, portability, etc.

So, the one real linux advantage just doesn't apply any more. The only single Linux box I have on my network is to support a third party application that the vendor has only certified on either RHEL/CENTOS, Solaris, AIX or Windows.

Which brings up another point - don't get into religious reasons for pushing an OS. If you have apps that are certified on OS Foo, then don't try and run them on OS Bar in production unless you have a very good reason to do so. The OS is an application platform - if your application doesn't run then it doesn't matter how good the OS is.
 
Morte said:
My reasons for using FreeBSD really don't have much to do with performance. It performs well enough, I don't care if it's a few milliseconds faster in some benchmarks or slower in others. In the end this almost always comes down to system/software tuning to really get performance.

After getting shafted by Redhat 8/9x I swore I'd only go with community supported distributions. You can never be sure when a company decides that you're "no longer commercially viable" and drop you as they please. What I like most about FreeBSD is that it is stable and consistent. It has a well defined division between the system, and third party software. I've had problems with upgrades sure, but I've always been able to fix them. Sometimes from some pretty messed up states. MS Windows doesn't expose the guts enough to make that possible, nor does it provide you with the tools to do that kind of thing. Linux has the tools and also does expose the guts, but what about the documentation? Maybe it's my imagination but it's hard to find answers to Linux related questions, because many distros do things their own way, use different sets of cobbled together software to implement the system, or the documentation you find is severely out of date. I still can't get Gentoo to boot into single user mode, nor can I find the correct answer to get it to do so.

The more I've learned about FreeBSD, the more I've been able to leverage the system. Learning Windows always made me feel like learning the latest techniques to apply band-aids and put up with the latest "great MS initiative", while learning Linux often feels like I'm learning the latest in fashion, only to be dropped when the next shiny idea looks better.

FreeBSD isn't the best for everything. I wouldn't go that way for a desktop, but for a server it's always my preference.

Exactly what I would say! :)
 
Back
Top