Handle With Linux

In response to my article:

10+1 things to tell your boss why you should migrate to Linux

An Article was written by Mike Dailey Titled:

Think before you speak migrating the enterprise to linux

criticism

While I welcome the criticism, I think it's to bad the author has taken it upon himself to try to take down all arguments. This pushed him to the point where he had to come up with arguments which can't be taken completely serious. This harms him, as I've even seen the term FUD been coined up against him.

While I don't think he is really against Linux adoption, I do think he went trough a little to much effort to make his statement. I think according to his conclusions I'll have to clarify some points in response.

Lower TCO

1. "Here the author makes and disproves his own argument; Linux does not deliver a lower TCO than Microsoft Windows server if the administrative costs are admittedly higher."

This is not what I stated, the overall cost of administration is lower, because less people do more work. This more than fully compensates for the increased cost per administrator. Doesn't it give an indication if something is more expensive by the way? It has more value, because it is more efficient. So in my opinion this argument still stands.

Security

2. "The long-standing argument of Linux security superiority has one notable flaw–network operating system (NOS) security effectiveness is a result of the technical skill of the admin, not the NOS itself. Any Linux distro, given to an inept admin with lacking security skills, will be far less secure than an out-of-the-box Windows server platform."

This argument effectively states that any secure system can be made insecure by incapable administrators. Which is an argument for what? Stating that an out-of-the-box windows server platform, which has every service it's build for disabled by default if I remember well, is more secure than a Linux system made insecure by an incompetent user, is a flawed argument at best. There are a lot of good arguments for Linux security, there are some arguments for windows security too. While it is true you can harden any system reasonably well, I personally wouldn't bet my life (or my job for that matter) on a windows system.

If the author could come up with valuable arguments why windows is more secure I would be happy to discuss my arguments why I think Linux is more secure. But for the moment I can keep it to the point that windows has a legacy of being a single user system, and has a monolithic design which are both bases for most of the security flaws in it's design.

Performance

3. "if Microsoft Exchange is the enterprise standard for electronic mail, for example, how does the Linux performance superiority argument factor in to the email environment?"

Exchange is a groupware server, not a mailserver and there are very good groupware applications available for Linux. "Enterprise standard" referring to what? If you need the same functionality this could arguably be achieved with open source and be higher performing. If it would be the smartest approach to start by replacing your exchange server is questionable though. If the only thing you need is a "mail solution" there's really no argument for not using postfix, as this is definitely faster and more secure than any possible windows solution.

Stability

4. "In the enterprise, neither Linux nor Windows should be deployed in a "multi-purpose" role.  Multiple applications or application types are not commingled on a server platform as this is a recipe for downtime regardless of the server OS."

This might be the authors experience, I have seen many Linux servers running many applications on single servers, stable and without problems for years. This statement would indicate that even a small company should be buying separate servers for everything to run stable. Windows maybe.
This is more an argument for Linux than it is against it, as where I referred to virtualization this is exactly the great benefit. Linux has very advanced technologies for kernel level virtualization, which makes it possible to run several virtually separated environments on the same kernel. This provides security and stability with almost no overhead, while consolidating many different applications on a single server.

Open Standards

5. "As much as we may dislike it, the truth is that organizations care little about open standards and industry competition. "

If they are well informed, they do care. If you are in a position of informing organizations about the implications of open standard and they "don't care" you have not done your job. Industry competition is not important to an organization? Competition keeps prices down and quality up, if that's no argument for a company than how can they survive?

Replacing Windows

6. "However, the author is arguing for the complete replacement of Microsoft Windows with Linux and Open Source alternatives.  Whether at the desktop or in the data center this is not a viable or responsible approach in terms of business decision or direction. "

I don't see any argument here why this wouldn't be a viable approach for business direction. Just stating this is no argument on itself. While I never stated windows should be replaced on short term and all at once, I do think it's a very viable approach to change from a windows environment to a complete Linux and open source based environment over time. I'm not saying any organization should be migrating to Linux without careful assessment and planning. I do understand the opinion, which many people share, for a mixed environment. But you should remember if your goal is to keep things that way, there's always a possibility of microsoft changing the rules. They will eventually come up with another closed protocol to keep your systems from communicating.
I could come up with a lot more arguments against mixed environments as an end goal. If it would be your intermediate goal there are certainly very good arguments for it.

Customized applications

7. "IT managers around the world cringe at the thought of customized applications.  Few things in the enterprise cost more in terms of budgetary dollars and man hours than the development and upkeep of a customized application.  This is the primary reason that most organizations seek out packaged or "canned" applications that will fit the business need.  Even in highly customized applications such as SAP the majority of organizations utilize the pre-packaged modules where possible to avoid the costs associated with customization."

Which is another argument for open source, not against it. Many organizations have a need for customized applications, which are far cheaper and easier to maintain if the biggest part of development is shared between many.

Open Source

8. "Few Open Source applications are designed from the perspective and input of a user base, instead providing only the functionality and options the author/developer felt were useful."

This is a very amusing perception of open source, while open source is developed by users I don't see "the perspective and input of a user base" in the response by the public to Vista.
I do understand that I'm writing to much out of the position of a experienced Linux user, as the author does clearly not understand what I'm saying here.
The point is: If you need anything in a Linux environment, there's a big change it is available for free. No need to buy it, just download it from a central repository, and start using it. This saves a lot of money and time.

Virtualization

9. "Again revisiting the argument of using free, Open Source alternatives, few IT decision makers are willing to deploy business-critical applications and systems on a virtualization platform that has little industry exposure, adoption, and support.  While the virtualization platform itself may be "free", the potential cost in terms of lost productivity as a result of the failure of that system far outweigh the costs of an industry-tested commercial virtualization platform. "

I'm sorry to say this but this statement does nothing but expressing a lack of knowledge of the subject. "Little industry exposure, adoption, and support." The author must have been living under a rock for a decade or so.

Future Proof

10. " It is doubtful that many executives, if any, consider the possibility of Microsoft going out of business.  While it is true that the Linux code would still be available if large supporters such as IBM or RedHat were to disappear it is also due that the majority of contributing code, drivers, and hardware support would disappear, as well, as these large supporters are the primary source of these critical aspects of Linux development. "

When it comes to risk assessment there are two things which are important:
risk and consequence
While the risk of microsoft going out of business might seem small, imagine the consequence.
If you are in a position of informing management about risk, you should argue in the consequence of going out of business if this happens. This is why organizations keep their data stored in remote locations. There is a very small risk your whole company will burn down, and all your data be destroyed. But as consequence would be going out of business, care is taken to have backup data off-site.
Furthermore open source is not limited to one organization, this makes the risk a lot smaller, and if they would all disappear the source and the open protocols do guarantee everyone can pick it up again and everyone has access to his data.

Happy system administrator

11. "If the author views the emotional state of the system administrator as the key reason an enterprise should switch to Linux it is easy to understand why the move for a steady migration to Linux is not gaining traction.  The idea that an enterprise would base technical direction on the happiness of technical support personnel demonstrates the disconnect between the Linux community and the realities of the business world."

If the author completely lacks a sense of humor, he still could have made up intelligently from the title which says 10+1 reasons, not 11, this reason had to be taken somewhat lightly.
This response clearly demonstrates the disconnect between the author and the humorous people of the world.

Comments

I don't understand how the comments to the authors response to my article can speak of a balanced view, which it clearly is not. While there are very good argument one can think of against open source and Linux migration, I don't think the author has stated any that could keep up against this list.

My Opinion

Any decision about any IT migration should be taken carefully, and all arguments pro and contra should be weighed in. There is no way one could ever say based on zero information if it is a good decision to start a migrating path to Linux. The article was only mentioning a few arguments for Linux adoption, and therefore it wasn't balanced by definition. If it would have been a balanced list of arguments for or against Linux in the enterprise, it would have been named different, and it would probably have been a list a mile long.