Jump to content

英文维基 | 中文维基 | 日文维基 | 草榴社区

Wikipedia:Reference desk/Archives/Computing/2012 January 24

From Wikipedia, the free encyclopedia
Computing desk
< January 23 << Dec | January | Feb >> January 25 >
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.


January 24

[edit]

One computer, several network connections

[edit]

Can you connect a computer (running Linux or Windows) to several Wlans and wired networks at once? Since you can add different wireless adaptors (through USB, for example), a shortcoming from the part of the network cards should be an impediment. 88.8.69.246 (talk) 15:04, 24 January 2012 (UTC)[reply]

The simple answer is that yes, you can connect to multiple networks at once, assuming you have multiple NICs (e.g. one wired, one wireless, giving two connections). However, actually using them simultaneously can be difficult. The solution will depend on what you wish to use them for. Are you trying to connect to the internet through multiple networks (in the past I have been on a flaky WiFi network, and used a 3G dongle as backup), or are you trying to stream home media through a wired network whilst internetting through WiFi, for example? There are hardware and software solutions that can be used, some of which get into the realms of 'pretty damn technical'. - Cucumber Mike (talk) 15:41, 24 January 2012 (UTC)[reply]
I was trying to use two wireless networks at once with a laptop. It has its wlan embedded, and I have a wifi pen-driver adapter. One network is the work network (so, no torrent download) and the other is a free network of a coffee place (which is perfect for downloading private torrents). Just trying to connect to both wlan doesn't work, although each one works fine for itself not simultaneously. 88.8.69.246 (talk) 15:53, 24 January 2012 (UTC)[reply]
Some people here and here are discussing this. The consensus is that it is difficult if not impossible, and the only feasible way is to virtualize an OS and present one of the adapters virtually to the OS --rocketrye12 talk/contribs 16:30, 24 January 2012 (UTC)[reply]
It's neither impossible nor very difficult. Connecting two networks turns your linux computer into a network device called a "router" or "network bridge". You may be familiar with such devices. The difficulty is that somebody - you, or your network administrator - must decide which traffic should route to which network: hence, routing.
If you'd like an overview, I found "An Introduction to Networking" in the NetBSD manual/user-guide. It contains a good explanation of multiple addresses, and the implications therein. Almost every tool discussed in this guide has a commonly-available free software version, and is generally available on most distributed variants of Linux.
On Windows platforms, particularly on the desktop operating system versions, "Network Bridging" is available, which is similar to general purpose internet protocol routing. Nimur (talk) 17:56, 24 January 2012 (UTC)[reply]
(Specific instructions: 24.6. Setting up a network bridge device, and 24.5. Setting up an Internet gateway with IPNAT). Nimur (talk) 18:25, 24 January 2012 (UTC)[reply]
I used to do this with Microsoft Virtual PC and 2 NICs and 2 CAT5 cables, very easy stuff. Von Restorff (talk) 05:33, 25 January 2012 (UTC)[reply]

"Unblowing" eFuses using a focused ion beam

[edit]

I've read that fuses like those found in embedded processors can be reconnected using a focused ion beam. Is this expensive? Does it work? --Melab±1 21:33, 24 January 2012 (UTC)[reply]

I'm guessing the cost depends greatly on whether you already have access to a focused ion beam. Since you probably wouldn't need to ask if you did have access, then I expect it would be expensive to purchase commercial access (if anyone does that) or worse buy a machine for that purpose. It may not be so expensive if you can convince someone to give you access at a research institution, which you could potentially do if you have some research that requires it, particularly if you have the backing of an institution or are a known researcher. But if all you're trying to do is to flash a locked Motorola phone then I suspect you're likely to be ignored if you ask for access. Nil Einne (talk) 23:26, 24 January 2012 (UTC)[reply]
Not too long ago, somebody asked if they could forge metal by melting down fountain pens. Your question is very similar to that question. Yes, in principle - this can be done. In fact, somebody, somewhere, owns a machine that can accomplish this task. Can you use a focused ion beam to perform surgical semiconductor modification? Yes - and you can also melt down fountain pens to extract the trace precious metals from them - but come on. Let's be realistic. This isn't going to happen. You don't have the expertise, the money, or the equipment. The knowledge of how to perform such a procedure is so specialized, that even Wikipedia, which has an article on everything, doesn't even have ion beam epitaxy... (probably we should link that to molecular beam epitaxy). This process is so specialized, so prohibitively expensive, ... that it's not going to happen. Be patient, excel academically, and do some research in a materials science or electronics lab, and eventually you'll be able to work on some cool stuff. Unfortunately, it's a fact that certain activities, like semiconductor lithography and steel blast-furnace operation, just aren't hobby-projects for the technical enthusiast. They never will be. Nimur (talk) 23:37, 24 January 2012 (UTC)[reply]
And even if the actual modification was cheap and easy, the getting into a position to do it on a real device would itself be an expensive undertaking. First one would have to deencapsulate the semiconductor (in a very clean environment) Then one would have to build up a detailed understanding of the circuitry (either by optical or electron-microscope examination), which is particularly difficult if the chip is heavily layered. Assuming the desired microcircuitry is accessible (that is, assuming it's not buried under other stuff) one would then do the beam thing. One would then need to reinstate the insulation, the thermal bonding, and the ceramic encapsulation. Doing this kind of thing only makes sense when one is trying to penetrate a sensitive package, presumably for the purposes of crime or intelligence gathering. And the kind of chips that organised criminals and intelligence agencies will want to crack will have layered protection against such penetration (where tampering with the packaging destroys the chip and wipes its payload). 87.113.28.157 (talk) 01:44, 25 January 2012 (UTC)[reply]