Computing desk
< October 2	<< Sep \| October \| Nov >>	October 4 >

Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

October 3

Virtualization

Man, it seems every day people get more and more excited about virtual machines and the likes. My question is, My friends and I have a gag that we like to try and see how deeply of a virtualizing we can get (Virtualbox in parallel in virtualbox and etc.) And we were thinking if there could actually be a practical application of multiple virtual machines... We could think of none. If the machine is being virtualized it has its limits from the program and blah blah blah. Does anyone have any counterstatements I can push back at them? I'm not talking about a master server running a DHCP server, and beside that a squid server, beside that etc. I'm talking about a computer running a dhcp server running a virtual box machine and inside that is another thing.

66.216.163.92 (talk) 01:42, 3 October 2008 (UTC)[reply]

I can only see testing a virtualization environment as the only worthwhile example. Assume you are a VMWare developer. You may want to test running virtualbox on it just to ensure it works. A second example, which I don't see as very valid because I disagree with the basic assumption of usefulness of the virtualization is a situation such as having every computer in an office running a virtualization image (so the IT staff only has to configure one computer, not all of them). Then, assume you work there and you want to run Linux on your virtual Windows box. You would be running some virtual OS in your virtual Windows. Of course, you wouldn't do that anyway since you would obviously kill the virtual Windows after a few days there and already have your computer reformatted and reinstalled with Linux. -- kainaw ™ 01:50, 3 October 2008 (UTC)[reply]

One thing all the commercial virtualisation providers are interested in is providing a general market for generic virtualised computing, much like Amazon Elastic Compute Cloud. In that business model large enterprises would have large physical compute infrastructures (whether big iron like Suns and IBMs or clusters of more modest hardware) and they'd sell compute time on that wholesale. Between those wholesalers and the final users of compute power (small websites, university projects, individual users) there might be several layers of middlemen and retailers; they break stuff down to more affordable chunks, and also aggregate otherwise unrelated chunks (e.g. cheap chunks bought from anywhere that has idle time, geographically distributed chunks to allow quick response to distributed users). While making each layer of this compute-cloud economy a concentric layer of virtualisation is, from a technical standpoint, rather inefficient, it might nevertheless turn out to be the easy and pragmatic way to structure things for business reasons. Still, surely you'd only expect to be a few layers of virtualisation from the hardware (in even the most byzantine market structure), and the economic pressures of disintermediation should keep the depth down pretty low. -- Finlay McWalter | Talk 12:15, 3 October 2008 (UTC)[reply]

Depending on the environment(s) you chose, a Russian nested doll virtualization setup could assist in security isolation in either direction (top-down-- isolating something you trust, or bottom-up-- isolating something you definitely do NOT trust). Having a number of layered hypervisors, especially on different operating systems and of different programmatic backgrounds, would lend itself to a self-contained heterogeneous system that proved very resistant to "generic" vulnerabilities. Of course, the key to this is the security of the hypervisor and the code that ties one system to the next (such as the virtualized network stack), so you would need a deep understanding from top to bottom to assert that this is worth anything but a lot of CPU cycles and disk space. --66.195.232.121 (talk) 19:44, 3 October 2008 (UTC)[reply]

Forwarding in Firefox and IExplorer

Here at the office, we connect to the internet with Nortel Guesspass. That means that if we want to reach the outside world (and not only the intranet) we are prompted for our password through the browser. Unhappily, the page that prompts us don't work in IExplorer, but we don't know exactly why. This page has a self-made certificate (not recognized by any authority). In Firefox we are able to add an exception to accept this certificate, however, IExplorer do not let's us add an exception and so we are not able to login using IE.

Any solution? Mr.K. (talk) 10:17, 3 October 2008 (UTC)[reply]

Umm... Request IT to get a legitimate certificate and in the meantime, use Mozilla Firefox? Is there something specific that you are unable to accomplish in Mozilla Firefox? I would assume that once you are connected to the Internet, you can use any browser you want. Kushal (talk) 21:23, 3 October 2008 (UTC)[reply]

Since the certificate is self-signed, I presume what you are seeing is the standard "invalid certificate" page introduced in Internet Explorer 7 and Firefox 3, respectively. As you say, the Firefox 3 one requires you to click through various dialogues to "add an exception" for the certificate.

In IE 7, it's actually much simpler: you are given 2 links, one of which says "Continue to this website (not recommended)"; clicking that link should load the page as normal.

I'm not sure if there is a way of permanently accepting an invalid certificate in IE7 (the above remembers your choice only for the current browsing session), but it looks like you may be able to add the certificate to IE's "certificate store": Once you're on the secure page, you should see a red "Certificate Error" box next to the address bar; clicking this will show extra information, including a link labelled "View Certificates"; from there, you can launch the "Import certificate wizard". Note that I don't actually know if this will help, as the test case I have is for a mismatched domain, rather than an invalid issuer, so importing won't help.

As Kushal says, it may be worth requesting your company acquire a legitimately signed certificate, but if this is purely an internal system, they may decide it's not worth the registration fees. - IMSoP (talk) 00:01, 5 October 2008 (UTC)[reply]

How about getting a real proxy server and dispense with this nonsense? ---J.S (T/C/WRE) 01:53, 6 October 2008 (UTC)[reply]

May not be a better solution, but how about making each person log in to the computer as themselves and then not restrict/authenticate Internet access at browser lever? Kushal (talk) 10:40, 6 October 2008 (UTC)[reply]

J.Smith: we thought Nortel would be the perfect solution.

Kushal: the problem with this solution is that we want to restrict how long people have access to the network. Visitors should not have more than 24 hours access.Mr.K. (talk) 11:49, 6 October 2008 (UTC)[reply]

Thanks, Mr. K. Just to be clear, everything still works as long as you log in with Mozilla Firefox, right? Do you NEED for your guests to be able to log in using Internet Explorer? If you are on a local wired network, and absolutely NEED to be able to use Internet Explorer, I'd suggest that you drop the certificate in its entirety. As a courtesy to Mozilla, I'd suggest you to use the "Report Broken Web site" in the help menu and let Mozilla know what you are unable to accomplish in Mozilla Firefox that you are compelled to use Internet Explorer. Cheers, Kushal (talk) 08:11, 7 October 2008 (UTC)[reply]

Actually, we don't need IExplorer. Nobody needs it. I use only Firefox, but some people are simply accustomed to IE and not planning to change. Dropping the certificate seems to be a good bet -easier than asking to change from IE to Firefox - since the problem is only caused by this certificate in combination with IE.Mr.K. (talk) 09:48, 7 October 2008 (UTC)[reply]

A lot of our faculty and staff complained when we went from MS Office 2003 to 2007. A lot of people complained against Facebook's new look (which is OK apart from being such a bandwidth hog and reportedly frustrating on dial-up). There are people who even complain against the awesome bar! Change is tough. I admire your sympathy with users accustomed to IE. Good luck (and I secretly hope you will be able to help some of your people to migrate to Mozilla Firefox)! Kushal (talk) 11:08, 7 October 2008 (UTC)[reply]

about array in VB

WHAT IS ARRAY? WHAT ARE THE VARIOUS TYPES OF ARRAYS IN VB? —Preceding unsigned comment added by 59.89.49.220 (talk) 11:07, 3 October 2008 (UTC)[reply]

Have a look at Array and this online tutorial. Feel free to come back here with further questions. Zain Ebrahim (talk) 11:10, 3 October 2008 (UTC)[reply]

Wrong Address on Google

Hello. I work for a non profit organisation which has a website listed on google. When you search for the name of our organisation, google's first hit is for our website - but the address and telephone number it lists is for a completely different business with a similar name. This has only happened recently - it's not from our site, it's the little map it shows at the top of google with the address and number. We've recently had a few calls from them saying that people have been calling them up when they're after us. What's the best way to get the address on google back to our address? Theorytest (talk) 12:40, 3 October 2008 (UTC)[reply]

It looks like Google provides some help on that question at their page here. Laenir (talk) 12:47, 3 October 2008 (UTC)[reply]

Perhaps your organisation's website doesn't have the address on it, in an accepted format that Google can recognise. -- Finlay McWalter | Talk 12:53, 3 October 2008 (UTC)[reply]

Just curious about website domain names

Why do you have to pay to get a domain name? For example if I want to create a website dhgfuigdagdhf.com, will I have to pay for the domain name? I don't believe anyone else is using this website right now, so isn't the domain name my own creation? Abhishek ^Talk 15:11, 3 October 2008 (UTC)[reply]

You're paying to be listed in the indices of the various DNS root nameservers. Some of these are operated by private entities, others by government bodies. None give anything away for free. -- Finlay McWalter | Talk 15:15, 3 October 2008 (UTC)[reply]

and (hopefully not to belabour the point) you can indeed create a website (say on a web hosting service like Dreamhost) without binding it to any domain. There it'll sit in splendid isolation, untroubled by search engines or other visitors. -- Finlay McWalter | Talk 15:37, 3 October 2008 (UTC)[reply]

Also, most domain names that people come up with are not unique which is why they have to try again. Who's to settle the argument over whether you or the guy who lives down the street came up with "berzerksheep.com" first? That's part of what registration does. --- OtherDave (talk) 16:04, 3 October 2008 (UTC)[reply]

It isn't the root nameservers but the nameserver(s) for the top-level domain, in this case .com. I suppose one can't expect them to give away the service for free, but the actual cost of each new domain to VeriSign is minuscule, probably less than $0.01 per year. -- BenRG (talk) 16:22, 3 October 2008 (UTC)[reply]

If domain names were free I would write a program to acquire every last one of them. My program would fail because someone who types faster than me would have beaten me to it. APL (talk) 18:46, 3 October 2008 (UTC)[reply]

That used to be the case. Originally, someone thought humans were nice trustworthy creatures that would only register domain names they planned to make good use of. Once domain hoarding started, the cost of registering domain names went up and up and up until it wasn't worthwhile to register every domain name that exists. -- kainaw ™ 19:11, 3 October 2008 (UTC)[reply]

Really it was a pretty stupid way to set it up from the beginning. Five minutes of consulting would have cleared up the idea that cheap domain names would lead to anything but wild irresponsibility. --98.217.8.46 (talk) 22:42, 3 October 2008 (UTC)[reply]

It depends on how jaded you are. You could be of the opinion that they did consult with lawyers. The lawyers discussed it and realized that if they allowed people to register tons of domain names, others would sue to get ownership of those domain names. A whole new area of law would emerge. The lawyers would make more money. So, they came back and suggested making domain name registration free - just to see what happens. -- kainaw ™ 23:23, 3 October 2008 (UTC)[reply]

To be fair, the Internet wasn't really open to "the "public" back then. It was all military and universities. APL (talk) 03:14, 4 October 2008 (UTC)[reply]

Isn't that like saying that opening an encyclopedia to editing by anyone would lead to wild irresponsibility? Anyway, as APL said, things were different back then. Commercial activity used to be banned on the Internet, because the universities and US government sites that funded it didn't want to subsidize for-profit use. And "ordinary people" hadn't even heard of the Internet; there was no incentive to register a domain unless you had a use for it.

Also, even at $0.01 per domain per year—heck, even at $0—it's not economical to register every possible domain name. There are 609,269,436,886,430,207,415,724,313,935,118,185,567,366,503,082,897,299,581,429,354,820,868,365,318,591,594,476,323,925,066,482,884 of them. Most of those are pretty long, but even a lower length limit of 8 characters, say, would be enough to effectively discourage people from registering random combinations of characters. They'd still register all the dictionary words, but they do that anyway. -- BenRG (talk) 08:54, 4 October 2008 (UTC)[reply]

No, it's not like saying that an encyclopedia open to the public would lead to irresponsibility—it's easy to reverse misuse of Wikipedia and it never costs anybody anything. It's not easy to reverse misuse of domain squatting and other domain ills. The issue is not with registering all combinations, obviously (straw man, much?), but with people selectively registering valuable/useful names and then hoping someone else will want to actually use it for something useful and pay them more for it. What did the squatter do to deserve that pay bump? Nothing. They didn't do anything creative, they didn't add anything to the world, and they didn't invest any real labor or resources into it. They just beat someone to the draw. I consider that a fatally flawed system. If it were up to me squatting would be disallowed by ICANN and challenges to squatting would be easy to file. --98.217.8.46 (talk) 04:09, 5 October 2008 (UTC)[reply]

There are already rules about squatting. If you own the trademark to a name and you owned it before the squatter, then it's no contest. There have been several court cases about this exact issue. ---J.S (T/C/WRE) 01:52, 6 October 2008 (UTC)[reply]

The trouble with that is that trademark law allows multiple companies to use the same name providing they are in different businesses. The classic (and L-O-N-G running) battle between Apple Records (founded by The Beatles) and Apple Computers is a good example of that. But sadly, apple.com can't belong to both of them - so the ownership of a particular trademark STILL doesn't guarantee you access to the domain name of your dreams. Seeing this coming, I registered a domain in my son's name when he was just a baby because I presume that having your own name as your domain is going to become increasingly impossible as time goes on. However, even then, I had to get him a ".org" address rather than a ".com" because of the expense (I went with oliverbaker.org - and it cost me a fortune even then!) That's really the flaw in the system. If the name is available - and I'm entitled to purchase it, it shouldn't matter whether the name is some gibberish like sdfaswdqwqq.net (very cheap) or something recognisable like stevebaker.com which went on sale for a couple of thousand dollars a few years ago (which was WAY too expensive for me!) - and what a useless site it actually is! I find that grossly unfair - the administration of the name servers, etc is identical for a gibberish name or a "good" one. SteveBaker (talk) 03:57, 9 October 2008 (UTC)[reply]

The actual operating cost for a domain name isn't just the bandwidth or the hardware... it also includes overhead costs for the entire business set up around selling them. Those things get expensive fast. ---J.S (T/C/WRE) 01:50, 6 October 2008 (UTC)[reply]