Linux: Look before you Leap

juniorboy sent us JP Mogenthal's latest piece from Internetwork, this timing cautioning people about Linux and along the way tries to get at the root of why people are switching.

Comments

Publishing this is illegal in America

[Anonymous Coward]

Almost every statement the author makes about linux in this article is incorrect. It's libelous and shows a lack of understanding about the subject matter.

Linux is an open-source project; therefore, .... Companies that add features they need, .... may find themselves in a redevelopment and retesting cycle every time a new version of Linux is released.

How is this different from other operating systems? With very few exceptions (such as afs) most programs only need to be revised between major version releases (if at all), and it was 3 years between 2.0 and 2.2.

With Linux, [internet] services will soon be available as a multivendor product. ... The most critical of these integrations will be security and access control.

Last time I checked, apache, sendmail, mysql, telnet, and ftp daemons were all included in most standard linux distributions. The linux kernel most definitely supports firewalling. All security holes that are found in any of these are patched quickly and an alert sysadmin has no problem keeping up with these updates. Microsoft patches to security problems (of which there are no fewer than with gnu programs) are much slower in general and generally are more harmful/widespread.

Without robust SMP, Linux servers can support only small companies and single applications. If you're managing multiple servers for increased scalability, you're better off using multiple NT servers all participating within the same domain.

2.2 supports SMP quite nicely, a lot more efficiently than NT deals with multiple processors. Neglecting Beowulf because it is still rather unstable and clunky, it is still quite easy to run linux machines in a distributed environment. You can divide UNIX servers into much finer granularity than NT as far as what services to run, and at a much lower cost. It also takes much less of a machine to run most services under linux than under NT.

Linux is a college student's project gone astray. ... Hey, I'm all for a competitor to Windows, just give me more than what Microsoft has to offer-not less.

Linus will agree with you that he did not intend linux to go this far, but it is not a "college student's project." Linux gives you far more on a server end than Microsoft has to offer. I definitely would not want to put a linux machine on the desk of every employee in management or H.R. or marketing, or even some engineers, but as a server operating system, NT does not offer anything better.

-- not anonymous, my company's firewall is just really annoying. nosilA (alison@andrew.cmu.edu)

Look Before You Leap

[dylan_-]

Look Before You Leap
Into Linux Adoption

JP MORGEN-THRALL
June 14, 1999

It seems I cannot read these days. I'm often seeing Linux touted as the greatest thing since sliced bread. Upon reading these rave reviews, I can't think. Most of this admiration is lost on me because I only know Microsoft Windows, as well as the relatively low IQ I have.

But I think it's critical that we, as a community, keep a perspective on the impact of Linux for the following reasons.

Linux is an open-source project; therefore, all changes are subject to review and approval by people. Companies that add features to the kernel, and then change to a different version will have to change the version, in much the same way as if you do something, you may actually do it.

Windows supporters still outnumber Linux supporters because I love pointy-clicky things. Windows NT Server Enterprise Edition has many fun things that I think come as standard because I never pay for them myself. I haven't a clue what comes with Linux, and I'm too lazy to find out, so I'll just make something up. Whereas Microsoft's products are buggy and barely usable, users may spend a significant amount of time trying to integrate these components under Linux, most of which I spent working out how to get the | symbol on this keyboard. The most critical of these integrations will be security and access control. I have no idea what this means, but it sounds impressive, yeah?

Linux is just beginning to be retrofitted for symmetric multiprocessing. Huh huh-huh...I said retrofitted! Without robust SMP, Linux servers can support only small companies and run a single application (honest!). If you're running NT you'll be required to manage multiple servers for increased downtime, and you're better off using an abacus.

Also, remember that Linux is still not Windows. One of the reasons for Windows' growth has been the complexity of obtaining a computer without it. Simply because a low-cost alternative is now available, it does not automatically generate more revenue for Microsoft.

Linux is a college student's project gone astray. There are pornographic pictures embedded in the kernel, and it drinks all your beer before throwing up on your brand new carpet. The version that will be supported by Sun Microsystems and IBM on its hardware will fall far short of each of these company's own Unix operating systems in marketing and hype. If you're responsible for operating system selection in your company, be wary of the Linux play. Hey, I'm all for a competitor to Windows, just make sure it's not a serious one.

JP Morgen-Thrall is president of NC.Focus, which provides strategic planning, analysis and psychic predictions of application integration technologies. He can be reached at jp@ncfocus.com.

dylan_-


--

I'm not even sure what the guy's point was

[yAm]

Windows supporters still outnumber Linux supporters because Microsoft provides a better value proposition. Windows NT Server Enterprise Edition ships with a full complement of Internet services, including Web, proxy, index, messaging, database, transaction and firewall services.

What?!?

For US$3500 you can get ten users using web, index and transaction services. Firewall? Perhaps he means the limited packet filtering that comes standard on WinNT. Database? I don't remember them throwing in SQL Server. Proxy 2.0 is anonther US$2000. Basically, Enterprise edition with all the stuff he mentioned runs:

• $3500 for NTE
• $5000 for Firewall (50 user Checkpoint FW-1)
• $2000 for proxy
• $1000 for SQL Server.
• $1500 for Exchange

$13,000 for an "integrated" WinNT Enterprise solution for 10 users...

Mayhaps he meant Windows Small Business Server. Still, I think the guy needs to see his dealer about the quality of his rock. It's giving him delusions...

Analysis or FUD?

[weatherwax]

On the surface, Morgenthal raises good points. These seem to be:

• It may not be possible for Corporation X to have its changes integrated into Linux
• Linux has fewer provided services and lower interoperability than NT
• SMP Linux lags SMP NT
• With it's Unix base, Linux is harder to administer than NT

At the risk of reiterating what most Slashdotters already know, I'd like to comment on this. BTW, none of the below is meant as a slight against FreeBSD or any other open source OS. Linux is the open source environment with which I'm most familiar - and of course, the focus of Morgenthal's article.

1. I'm not aware of any process whereby a corporation can have its changes integrated into NT. Nor any other commercial OS, for that matter.

   On the other hand, with Linux,

   • such changes are possible, and if they make sense to the project at large, will be included
   • modules allow the OS to be extended without kernel modification

2. Linux services are certainly more loosely coupled than those from Microsoft. I can't off-hand think of anything which is missing from a Linux distribution, including a choice of database servers, but it's true that for a particular application, it might be necessary to work with a number of third-party components to meet the same level of integration.

   On the other hand, once this integration is achieved, it will not be dependent on proprietary protocols (such as Exchange), will be upgradeable at the component level, will not be as susceptible to email attacks, will be more secure and more stable. Etcetera. In a lot of integrations, these would be seen as advantages.

   Also, of course, the level of interoperability and integration is increasing exponentially, particularly with Gnome, KDE2, CORBA-compliant applications - and this integration is happening with multiple vendor support.

3. Linux SMP, except in some narrow benchmarks, has held its own against NT for some while. Until recently, multiprocessor NT systems have been barely more effective than a single processor version. At the same time, Linux SMP has been becoming more efficient; it may be that NT is currently outperforming Linux SMP, though I suspect this theory is largely based on the Mindcraft test results. However, as the 2.2 kernel is being improved, that situation isn't likely to last.

   At the same time, of course, NT scales beautifully -- to a bigger, faster, Intel processor, while Linux scales to faster machines and upcoming processors. SMP addresses performance, not integration, and in this area, Pentium-specific NT isn't likely to maintain a lead.

4. I would submit that someone who claims NT is easier to administer than Unix has not spent sufficient time learning one or the other environment. NT isn't easier, it's prettier.

   Going beyond basic configuration becomes very hard very quickly with NT, as the menu/dialog-driven utilities allow limited selections, and have limited debugging options. In fact, one of the most useful tools for administering an NT network is a Linux box with Samba and tcpdump.

   A medium sized company's networking needs may be able to be met by NT, but configuration management requires every bit as much of a networking guru as the equivalent *n*x network, and troubleshooting can be far harder.

   I'm puzzled by the comments of Linux's inferiority to the Big Boys' Unixes - Linux ease of configuration and use seems to compare very favorably to AIX, HP-UX, etc.

5. Finally, Morgenthal makes an off-hand comment which I believe speaks to his motivation. Linux certainly began as s student's project. To speak of it "going astray" is to make a perjorative comment about its rate of increase. "Becoming too big to be manageable" might have been a fair comment, though I'd disagree with it. But to speak of an OS which is already extremely reliable, well-supported and widely used as "going astray" reveals a personal bias against Linux which at least would give an impression that the article is designed as FUD rather than an impartial analysis.