Systems Manageability - Part 1: Why Manageability Matters

Systems Manageability – Part 2: Scope, Methodology and Ontology

Port 25 — Fri, 06 Apr 2007 21:27:26 GMT

In my last blog called “Why Manageability Matters” I talked about why we chose to work on “Systems Manageability” as a whole and get a grassroots understanding of it within the context of Linux and Open Source space. In this blog, I’m going to address

re: Systems Manageability - Part 1: Why Manageability Matters

ajaymm — Fri, 23 Mar 2007 23:03:40 GMT

Hi Kishi,

Thanks for doing a great job in your blog dealing with Manageability. I have subscribed to your feed, so I will watch for your posts on this topic. I would love to connect with you and talk more about what we are doing. We are also looking at a lot of open source initiatives around manageability and linux. Please send me an email - ajay.mungara@intel.com and we can sync up.

Also, check out my blog at http://www.intel.com/software/ajay

re: Systems Manageability - Part 1: Why Manageability Matters

kishi — Fri, 23 Mar 2007 19:28:44 GMT

Ajay - thanks so much for posting a comment, this is very helpful. Over the next seven or eight more blogs, I will try to drill down on the ontology of how I saw and experienced manageability being approached in the Linux and open source space and I hope that sheds light on a part of what you've brought up. On a personal note, would be very interested in talking to you about your role as the manageability developer community manager for Intel. Let me know how and when we can connect.

re: Systems Manageability - Part 1: Why Manageability Matters

kishi — Fri, 23 Mar 2007 19:24:53 GMT

einhverfr - first off, thank you for your comments. I agree with what you're saying about Heterogenous manageability and why having a diverse set of platforms is what the challenge is around. And yes, Linux systems need to be maanged differently than Windows systems. However, for a large organization - that's where they need the most relief and which is why you pay more for a framework ultimately, than you do for an application

re: Systems Manageability - Part 1: Why Manageability Matters

ajaymm — Fri, 23 Mar 2007 16:50:14 GMT

Hi Kish,

You are right on. Manageability is really central to the success of any IT department. There is huge a tendency to simplify the environment to offset the risk associated with manageability and rightly so. But, the end-user demands on technology have risen sharply and they want the latest and greatest tools. Striking that balance between manageability and the user demands will the key to success. I would like to see your take on infrastructure security in association with manageability. I think you are not going to have much argument around “why manageability maters” because it has been consistently one of the top 3 concerns/challenges of the IT departments for many years. But, the key would be to understand “how to address the manageability challenge in a complex, ever changing, and highly demanding IT environments”. I work for Intel Corp. as a manageability developer community manager. We are trying to address many of the manageability challenges with Intel® Active Management Technology and I am always looking for new ideas around how to address manageability.

Regards,

Ajay Mungara

Software Solutions Group

Intel Corporation.

re: Systems Manageability - Part 1: Why Manageability Matters

einhverfr — Thu, 22 Mar 2007 02:20:34 GMT

One of the issues of manageablility that occurs to me is that heterogeneous networks have to deal not only with the fact that different software runs on different systems, but also with the idea that the platforms themselves were designed around very different environments.

Active Directory/GPO is the ideal management framework for Windows, while something like Athena (though perhaps with diskless workstations) is the ideal management framework for Linux.

If you deploy Linux like you would ideally deploy Windows, the support and manageability issues become quite problematic for larger networks.

But you can't really deploy Windows the way you would ideally deploy Linux and have it be very useful. I suppose if you tried hard enough you *could* but again, the support and management would become a nightmare for anything more complex than a very simple network.

So, in my view, a heterogeneous network means that you often have two very different networks which are sort of bridged together hopefully with a framework of some common tools. Unfortunately, this rarely happens though. (IIRC, MIT does this. though.)

So one of the issues I have never really seen covered is the mismatch of the design criteria of platforms in the area of manageability. I don't know if this is likely to be covered this time around, and I suppose not enough businesses who run Linux take full advantage of it, but it might be something worth further consideration.