Lights out management - still waiting for the bulb to glow
Go vertical or go home
Blog I am on a quest to lower my computer power usage. If you have read my previous article, you know that this is not by choice - it’s a necessity driven largely by cooling requirements in the spaces where my systems live. The project at hand is Lights out Management (LOM), the ability to configure and control my systems even when they are powered off, thus allowing me to keep them powered off unless I need them.
To recap: I discovered with a little bit of hunting that I could do some really cool things. Wake On Lan (WOL) can wake sleeping, hibernating or suspended systems from their torpor, while Simple Network Management Protocol (SNMP) can give me granular control over my uninterruptible power supplies (UPSes), power distribution units (PDUs) and various flavours of printer. When it comes to actually getting some out-of-band remote management on these systems (in case they experience an error while coming back online) I have at my disposal an IPKVM and several systems running Intel’s vPro.
The goal is to actually find some software that allows me to use all of these tools from one interface, schedule power downs and wake ups and poke at my systems to see what is wrong with them if they aren’t responding. Preferably, I need to come up with something that I can throw together in the next few days, learn to use in no time flat and only poke at when there’s an actual problem needing my attention. While I normally love tinkering with projects endlessly, time is short and this LOM project needs to “just work”.
Whether open source, proprietary, free or paid, there are many enterprise desktop management tools out there capable of some aspect of LOM control. I have two requirements: the first is to assemble a set of basic tools for manipulating all elements of my LOM setup independently. The second requirement is to find a centralised management tool that will allow me to run my LOM setup across my whole network seamlessly and easily.
The search for candidates to include not only in this article, but to actively deploy on my networks left me nearly weeping with frustration. My Google-fu was inadequate, because if there exist applications that actually take LOM seriously I had virtually no luck finding them. Everything I could find that appeared to begin to take it seriously was part of some vertical stack of applications offered by OEMs like IBM or HP. Third-party, platform-independent lights out management is thin on the ground. To achieve the first requirement of my project, I need to acquire some backup management tools. These have to be simple tools, or preferably a single tool that does all the things I needed simply.
vPro and WOL proved to be the easiest of the lot. Applications that can reach out and wake a computer up using WOL number in the thousands - pick your poison. I will stick with what I know, and so for my “simple WOL tool” I have a combination of the WOL abilities built into my DD-WRT Wi-Fi routers and a tiny Windows app called “magic packet sender”.
Intel offers a Manageability Developer Toolkit which comes with a program called the Manageability Commander. While I am still not 100 per cent certain if the licensing terms allow me to actually use this in a corporate setting as a regular tool, if you have a vPro system and want to start making use of it right now, this is what you were looking for.
I could find two out of the three tools to satisfy my requirements, but APC stubbornly refused to play ball. I am starting to believe that it is quite possible there is no such thing as a simple anything when it comes to managing APC equipment. APC offers centralised management software of its own, but it is anything but simple and the costs are exorbitant, even for APC. For the cost of the software, I could replace every piece of APC hardware I have with equipment from an alternate vendor that would play nice with various bits of open source software I have grown to like.
"I do maintain that the goals that open source communities pursue (making software that does what they want) and commercial vendors pursue (making money) are quite different, and that assuming they're the same doesn't make for a meaningful comparison."
If we were having this debate ten years ago, I would have agreed with you wholeheartedly. Times have changed however, and I feel your statement reflects a small minority of extant open source contributors. Open source software is about business now, just as closed source software is. I am sure there are the odd programmers out there who contribute a bit of code for the fluffy bunny cause. Perhaps they wrote something cool and they wanted the world to be able to share in it. (I’ve written a few of those myself.)
These are more and more occurring in the minority. Open source has become nothing more than another element in various business models. Like all of IT, regardless of the original intentions or philosophy, the religion was jettisoned and it is now about nothing more than business.
As to the community, be it proprietary or open source having failed; I maintain that they have. They have failed to see the requirement for a LOM network management application and create one.
Again; your assertion is that it is my duty as a systems administrator to get custom code done for anything I desire; a very programmer-centric view of the universe. I have a sysadmin-centric view of the universe: it is the duty of the programming community to have software for every conceivable need already available “on the shelf.” Preferably more than one option, so that competition can drive refinement of the applications and expansion of features.
You and I will probably never be able to agree on this, because we have different views of whose responsibility it is to recognise the need. As a sysadmin I say that it is the responsibility of entrepreneurs and developers to see a need, take the risks and write an application. This then gives them the intellectual property rights to the application, either to release it under a proprietary or open source license, and to recoup expenses via whatever business model they desire.
If the onus is on my to identify and then commissions these works, then the intellectual property belongs to me, as the individual who commissioned the work, and it would be lunacy to ever release that software into the wild. By commissioning it to meet the specific purposes of my business, it’s existence provides me with an advantage I have over my competitors. No business man is going to give away their edge.
Thus in the case of most businesses, if they follow your advice, they get the software they want, but that software is never released into the wild. Good for the company, (they get what they want,) good for the developers, (they are paid to develop it,) but bad for anyone else. The burden of risk here is on the individual business commissioning the work, and they would be the sole beneficiaries.
In my case, the developers take on the risk of funding the project based on the idea that they will be able to recoup expenses by selling their application. (Or support for their application, or what-have-you.) If the developers are good, and market savvy, it works out well for them. If not, they lose. If they write a good product, then all of their customers benefit, the developers benefit, everyone wins.
That this particular niche hasn’t been identified and exploited I continue to see as a failure. Maybe some companies have commissioned custom apps. For obvious reasons they never made it into the wild. Others have been bought up by the large OEMs. Lacking competition from third parties, the large OEMs can charge exorbitant rates for software that does what I am seeking. (Worse yet, they tie it specifically to their hardware and software stacks in an attempt at vertical integration.)
I am sorry the statement rubs you in the wrong way, sir…but my view on who “should” be assuming the risks, and who bears the responsibilities in this are based on the “rules” of the capitalist society I inhabit. (Not all that willingly, I might add. I am a socialist at heart, but paid to think like a capitalist as part of my job.)
In the meantime, I’ll just use Spiceworks.
failure through lack of crystal ball
Interesting take. I'll buy your assertion that you ment it impartially, but I'll observe that I read it differently and you could've worded it so that I would've read it how you ment it the first time. I'll also observe that you don't quite as often see assertions "industry has failed to provide" when reviewing a number of commercial applications where none of them do exactly what the reviewer was looking for.
I do maintain that the goals that open source communities pursue (making software that does what they want) and commercial vendors pursue (making money) are quite different, and that assuming they're the same doesn't make for a meaningful comparison.
If you do insist on treating open source like you would a commercial entity, then make the comparison fair and hire an open source consultancy that can tailor the application so that it does what you want it to do. You can then compare it on equal or at least comparable footing and you get the warm fuzzy feeling of having contributed back to the community too.
In neither case does it make much sense to claim other people fail, unless you had a contract with them that specified they perform and you can demonstrate they didn't, or didn't completely. You failed in your pursuit, and you found a niche. That's an opportunity that someone, possibly you, might take up to fill the niche. And that, bringing something to market that wasn't there before, is innovation.
Getting people to fill the niche requires getting the message out at the very least, and we're back at communication again. If you think asking for features from open source communities is too much like following evangelism, well, you could call a sales agent and ask for the feature in a commercial application. A good sales agent will be happy with the question and pass it on to the developers. Because they, too, need to know what the client wants.
If no corporations offer the product I require, then the corporate community has indeed failed to deliver. "The open source community" doesn't get a free pass just because it's open source. If I went to a car dealership and required a car that had both anti-lock breaks and an MP3 player, I would consider it a failure of car manufacturers to anticipate demand if there were no models available with both options.
I don’t care if “the open source community” has no single CEO or board of directors. “The open source community” gets no slack, special favours or happy meal points from me simply because of their business model. If it wants to think of itself either as a bunch of individuals with no cohesive direction, or a hivemind monstrosity there is no difference to me. “The open source community” in whatever format he/she/it/they/pie chooses to look at itself as simply hasn’t produced the right shaped peg for this particular hole.
It is not up to the implementers of software to adapt to the whims of developers; it is up to developers to anticipate the needs of implementers. In this case, developers have failed to recognise and exploit an open niche. To me that is a (to date) failure. You can disagree with that all you like, or try to argue semantics or terminology, but as far as I am concerned it simply is a failure to recognise need and opportunity.
You choose to take my words as complaint; they aren't. They are merely statement of fact. In the case of "delivering what I require for desktop management software," the open source community has failed. For that matter, the paid community has largely failed as well. As good as Spiceworks is, there are some holes in it that need work.
Pointing out gaps in offerings isn’t complaining. It’s merely being observant. There’s also dash of holding everyone, regardless of business model or software licensing terms, to a high standard.
You can of course take the typical open source evangelists’ approach of “well, if there isn’t the application/script/etc. that you want, then you write it and quite expecting someone else to.” I’m not a developer by trade; I’m a sysadmin. I can script, but I certainly can’t write anything the quality of Nagios or Spiceworks. I also shouldn’t be expected to have to be a developer of that calibre just to maintain a network. “IT” is so large that it requires many different people of many different skill sets to keep operational. My job is to find the best of what is available and put it to use. It is the job of other people to identify software niches and fill them.
Those people have in this case failed. With luck, a group of developers is right now working on a product to fill exactly this gap; if so, I applaud them and look forward to trying their software out.