MIT Bandwidth

<p>You're talkin of who invented the internet and NO ONE has spoken of Sir Tim Berners-Lee?? He invented: www, html, url, hyperlinks, web browser...</p>

<p>He's currently at MIT.</p>

<p>Sir Tim Berners Lee = STBL</p>

<p>Anyway, so STBL is annonoucing that he is developing Internet 2.0. It will really help with my The Research Project, though I am betting everything I own, as soon as I release a beta, M$ will find a way to shove my code up it's @$$ and sell it... :)</p>

<p>I don't think he is a particular genius though compared to some of his colleagues. His biggest role was developing the interfaces, and standardizing them. The internet was developed by Gov and Uni's(MIT is the biggest).</p>

<p>BTW. Do you guys consider Ada Lovelace to be the first programmer?</p>

<p>Yeah, Berners-Lee is at MIT right now. But when he invented hyperlinking and www, he was at CERN or at ICS. Hence, I don't know if it's correct to say that MIT has a lot of IP addresses because Berners-Lee happens to recently be affiliated with MIT. </p>

<p>I would also ask why some of you keep insisting that MIT is the biggest uni in developing the Internet. I don't think there is any evidence to support this contention. MIT was one of the original 15, but was not one of the original 4 Arpanet nodes. Furthermore, with the possible exception of Kleinrock (who I see as really a prophet, rather than a true internet principal), it is difficult to say that MIT was the biggest developer of the Internet - other schools, particularly UCLA (the alma mater of Cerf, Postel, and many other early Internet luminaries) seem to hold a much stronger claim to that title.</p>

<p>Trust me, while MIT's network is good, it's not THAT good. While the pipes are big, what really matters is not the size of the pipes themselves but rather the size of the pipes relative to the load on them. Let's not forget that MIT is full of thousands of students and researchers, many if not most of them who are tech junkies and who tend to soak up huge amounts of bandwidth. While there are times (especially during the holidays) when the available bandwidth really is blazingly fast, there are other times when lots of people are on the network and consequently things get really really slow. There are times when students have said they had to leave campus and go to their DSL-enabled apartments in order to get decent Internet bandwidth because so many other people on campus were clogging the lines.</p>

<p>Which students would say that ,not to brag about my dl speed but i have reached up to 2 3 mb/s connections but i hear other places get slower connections , I never worry how fast my connection but how fast is the server sending .</p>

<p>In other related news, Robert Metcalfe, founder of 3Com and inventor of Ethernet was an MIT alumnus.</p>

<p>


</p>

<p>Nope, so far nothing. :(</p>

<p>2bad4u, have you ever tried downloading big files in a dorm at, oh, I don't know, say, the late afternoon, when the network tends to be impacted? </p>

<p>Look, I've worked at places that had a straight FastE drop right to the Internet, and I was regularly exceeding over 50Mbps+ of download speed, often times more. I can also say that the network speeds I can get at MIT are not appreciably better than what can be had at that 'other' school in town.</p>

<p>
[quote]
and I was regularly exceeding over 50Mbps+ of download speed

[/quote]

What u need that sort of bandwidth for? The random read speed of most hard drives is ~10 Mbps if i'm not mistaken.... and 20 Mbps happens to be the total bandwidth available to many ISPs here in india.</p>

<p>Give me 0.5 Mbps, e.g., and I'll be your slave for ever.</p>

<p>Let me put it to you this way, mercurysquad. Let's say that it is true that you can't usefully use more than 10Mbps of bandwidth. So if that's true, then why is it that Ethernet LAN cards on most systems have progressed from the the standard 10Mbps to 100 Mbps and now Gigabit speed? Why do NIC vendors ever bother to improve their cards if the extra bandwidth is never used anyway? Are they being dumb? And it's not just NIC cards that have to be changed out to improve bandwidth. The backend Ethernet switches also have to be upgraded to accomodate greater and greater ethernet speeds, and many US companies have dutifully upgraded their LAN's from 10 Mbps to 100Mbps, and nowadays, many are upgrading to GigE to the desktop. These upgrades take a lot of money - you gotta replace blades on the switch, sometimes you gotta get an entirely new switch, etc. Why are they doing that if their computer systems can't take advantage of all that bandwidth anyway? Are they just stupidly throwing their money away?</p>

<p>Mb/s --> megabit/s, 50 megabit/s = 6.25 megabyte/s. very fast for internet, but def. not higher than the top hdd read/write speed. if you want data, look at <a href="http://www.dslwebserver.com/main/fr_index.html?/main/ata33-vs-ata133.html%5B/url%5D"&gt;http://www.dslwebserver.com/main/fr_index.html?/main/ata33-vs-ata133.html&lt;/a>. 86.0MB/s (MB - megabyte, NOT Mb) on ata/133 for example.</p>

<p>=)</p>

<p>Wow. I never said a lot of bandwidth was bad.. I just remarked that as an isolated example of a home/office user, why would anyone want 50 Mbit/s of bandwidth when, e.g. 1 or 5 might be more than enough. I'm on a 115k connection which rarely works at above 80 kbps. If I had a 500 kbps connection, I cannot picture myself complaining for more bandwidth.</p>

<p>And karthik, try running a random read test (e.g. SiSoft Sandra) on your average harddrive, and see if you ever get 86 MB/s (or even 20). Theoretical limits are great, but aren't always approached. ATA/133, for instance pushes the limit further, but at the end of the day your computer will find itself reading file fragments scattered on the entire harddrive, resulting in read or write speeds not significantly higher than ata100, which is the current defacto standard......</p>

<p>Oh, when I say 'your average harddrive' I mean - "an average harddrive" - not that your harddrive might be average :)</p>

<p>no worries =). just wanted to point out that the 50Mb/sec was far below the real limit for hdds (the site i link to posts actual benchmark tests, albeit with an <em>extremely</em> nice hdd).</p>

<p>sorry if i came across the wrong way! :)</p>

<p><a href="http://web.mit.edu/network/mit-net.pdf%5B/url%5D"&gt;http://web.mit.edu/network/mit-net.pdf&lt;/a&gt;&lt;/p>

<br>


<br>

<p>Wow.</p>

<p>On a semi-OT note, what types of connections are we talking - is the campus blanketed with wireless, or do you have to provide your own if you want it? I'm assuming the rooms have ethernet at least, right?</p>

<p>all rooms have an ethernet drop per occupant; many/most classrooms have outlets, also. campus is blanketed with pretty nice wireless, i hear, as my wireless card is flaky and doesn't always play nicely with the nodes.</p>

<p>i find the bandwidth to be sufficient, tho i'm sure out in the Real World nicer setups exist. but then, i'm insufficiently techy to really talk about these things.</p>

<p><a href="http://whereis.mit.edu/map-jpg?provider=0&provider=1&Wireless-provider=go%5B/url%5D"&gt;http://whereis.mit.edu/map-jpg?provider=0&provider=1&Wireless-provider=go&lt;/a&gt;&lt;/p>

<p>If you want REALLY fast HDD access, just build a solid-state HDD from RAM. :)</p>