February 1, 2007
One of the sadder recent news stories is the disappearance of Turing award-winning researcher Jim Gray. I've written about Jim's research before; he has a knack for explaining fundamental truths of computer architecture in uniquely clear ways.
For example, in this ACM interview, Jim illustrates how the unusual economics of bandwidth can make a sneakernet worthwhile – if you're sending a terabyte of data.
JG We built more than 20 of these boxes we call TeraScale SneakerNet boxes. Three of them are in circulation. We have a dozen doing TeraServer work; we have about eight in our lab for video archives, backups, and so on. It's real convenient to have 40 TB of storage to work with if you are a database guy. Remember the old days and the original eight-inch floppy disks? These are just much bigger.
DP "Sneaker net" was when you used your sneakers to transport data?
JG In the old days, sneaker net was the notion that you would pull out floppy disks, run across the room in your sneakers, and plug the floppy into another machine. This is just TeraScale SneakerNet. You write your terabytes onto this thing and ship it out to your pals. Some of our pals are extremely well connected – they are part of Internet 2, Virtual Business Networks (VBNs), and the Next Generation Internet (NGI). Even so, it takes them a long time to copy a gigabyte. Copy a terabyte? It takes them a very, very long time across the networks they have.
DP When they get a whole computer, don't they still have to copy?
JG Yes, but it runs around their fast LAN at gigabit speeds as opposed to the slower Internet. The Internet plans to be running at gigabit speeds, but if you experiment with your desktop now, I think you'll find that it runs at a megabyte a second or less.
DP Megabyte a second? We get almost 10 megabytes sustained here.
JG That translates to 40 gigabytes per hour and a terabyte per day. I tend to write a terabyte in about 8 to 10 hours locally. I can send it via UPS anywhere in the U.S. That turns out to be about seven megabytes per second.
DP How do you get to the 7-megabytes-per-second figure?
JG UPS takes 24 hours, and 9 hours at each end to do the copy.
DP Wouldn't it be a lot less hassle to use the Internet?
JG It's cheaper to send the machine. The phone bill, at the rate Microsoft pays, is about $1 per gigabyte sent and about $1 per gigabyte received – about $2,000 per terabyte. It's the same hassle for me whether I send it via the Internet or an overnight package with a computer. I have to copy the files to a server in any case. The extra step is putting the SneakerNet in a cardboard box and slapping a UPS label on it. I have gotten fairly good at that. Tape media is about $3,000 a terabyte. This media, in packaged SneakerNet form, is about $1,500 a terabyte.
Does transferring a terabyte of data via sneakernet make sense?
First, consider the bandwidth capabilities and monthly cost of a few common Internet connections.
Of course, costs may vary; I chose costs that jibed with my personal experience and lined up with a few cursory searches for pricing around the web. Please let me know you think these costs are way out of line. Assuming for the sake of argument that these are representative costs and throughput rates, how much would it cost to transfer, let's say, a 20 gigabyte high definition video file?
||Download 20 GB
||Upload 20 GB
And how much could we upload or download in total, assuming we had these connections going full-bore, around the clock?
||in 24 hours
||in 1 month
Let's say we wanted to send a terabyte of data via sneakernet:
- Two 500 GB hard drives weigh about five pounds; we can wrap the drives in bubble wrap and fit them in a standard FedEx box.
- It costs about $60 to send five pounds in a standard FedEx box coast-to-coast in 24 hours.
- The total transit time is 32 hours: 24 hours, plus 8 hours to copy the data on and off the drives.
We just transferred data at the rate of 9.1 megabytes per second. The only internet connection that's capable of our sneakernet throughput level is the OC-3. None of the others are even close, particularly if you consider the highly asymmetric nature of consumer connections, where upload rate is a fraction of the download rate.
And what about the cost? Not including the $300 expense of the two hard drives (which I think is fair, beause they're reusable), the total cost per gigabyte breaks down like so:
||Cost per GB
|Cost per GB
It wasn't obvious to me, but the sneakernet math clearly works. This is exactly the kind of insight Jim Gray was famous for.
Jim also says the cost of internet bandwidth was roughly a dollar a gigabyte for Microsoft in 2003. Is that still how much internet bandwidth costs today? According to the figures I found, the only connection that expensive today is a modem. And who uses modems any more? It seems implausible that consumer internet bandwidth would be sold cheaper than large blocks of commercial internet bandwidth. Let's take a look.
I'm not sure who to believe. It's a good sign that most estimates are under the $1.00 per gigabyte rate that Jim quoted in 2003. I'd like to think that the cost of internet bandwidth is getting less expensive over time. High bandwidth costs lead to a de-facto "popularity tax", and that's like a giant wet blanket over content creators. Cheaper bandwidth is a net public good: it leads directly to more content, and higher quality content, for everyone.
Posted by Jeff Atwood
Sneakernet absolutely makes sense, this is a problem I've run into in my own work. We occasionally had to ship several terabytes, sometimes we'd do it over the internet, sometimes by shipping hard drives. Your math makes perfect sense, but when you take into account the fact that your pipe usually has a heck of a lot more to do then just send that one job, it makes much, much more sense. One of my first real jobs in the industry was writing software to take our low-priority jobs, break it into pieces, and send it at night when the load was low, so it didn't inconvenience the other users.
I always thought it fairly amusing that the low-priority jobs got sent at night, over a period of days, whereas the high-priority stuff got stuffed in a box and given to the postman. It says a lot about the current state of networking, both how far its come (in years past, it would be a no-brainer. Like they say, never underestimate the bandwith of a station wagon full of mag tape) and how far we have yet to go.
As the old saying goes...
Never underestimate the bandwidth of a station wagon full of tapes driving down the freeway.
Jim Gray may well have been paraphrasing the well-known (in English circles anyway), "Never underestimate the bandwidth of a Ford Transit full of CDs". I guess you could scale that up to HDs quite easily :-)
P.S. Just don't expect a ping that you can use in CounterStrike.
That "old saying", I believe, is originally from Andrew Tanenbaum's "Computer Networks" and goes like this (I just looked it up in my Second Edition copy from 1989, page 57): "The moral of the story is: Never underestimate the bandwith of a station wagon full of tapes hurtling down the highway."
I still remember laughing out loud when I read that line for the first time in college. It was such an eye opener.
Plus, there is something to be said for the fidelity of sneakernets. When the guy in the brown shorts shows up with your package, the transfer is done and you know (or hope) the package is complete. ZMODEM, FTP, and HTTP can't touch that.
Indeed, computer networks only give you a better response time. You can establish connection in milliseconds, but when it comes to bandwidth, networks lose even theoretically when compared to physical movement of media. Networks use very narrow (in physical size) channels and very limited set of media states to transfer signals. Even with optical network from one end to another, you can use that same technology to fit an orders of magnitude more data to a 5"-sized media and then deliver it by mail. Not to mention cost of dedicated optical channel.
By the way, what about using airliners or ships loaded with DVDs for intercontinental networking? :)
You would easily run into scenarios, though, where you would want your empty HD shipped back. In which case you would have to include the cost of shipping back the two HD, thus nearly doubling the cost, though saving the 16hrs at either end for upload and download. This translates to a round trip total time of 56 hours, a total cost of $120, a data transfer rate of 5.08 MB/sec, and a cost of $0.12 per GB transfered.
I'm also not sure where you got your transfer rate figure of $0.08 per GB for Sneakernet. seems like it should be about $0.06 per GB.
where you would want your empty HD shipped back
Why? If they're empty, what am I getting back? But maybe you're right; if I don't get them back eventually, my real cost was $360 ($150 per drive, plus $60 fedex), which makes the transfer cost $0.36 per gigabyte. Ideally you'd get them back with some other data that person needed to send you.
I'm also not sure where you got your transfer rate figure of $0.08 per GB for Sneakernet
Not sure what I was doing there; the time taken isn't relevant, it's just the price divided by total size of data transferred. Corrected.
Networks use very narrow (in physical size) channels and very limited set of media states to transfer signals
That's right, and one of the biggest problems we're facing right now is the storage explosion -- we have mountains of cheaply stored data. That's not a bad thing. But the pipe to get to that data is growing very, very slowly. It's the disconnect between these two growth rates that's the problem.
One interesting thing happens as hard drive sizes increase, without any comparable increase in bandwidth to get to the disk: you have to treat them like sequential, tape-style devices.
JG Certainly we have to convert from random disk access to sequential access patterns. Disks will give you 200 accesses per second, so if you read a few kilobytes in each access, you're in the megabyte-per-second realm, and it will take a year to read a 20-terabyte disk.
If you go to sequential access of larger chunks of the disk, you will get 500 times more bandwidth—you can read or write the disk in a day. So programmers have to start thinking of the disk as a sequential device rather than a random access device.
DP So disks are not random access any more?
JG That's one of the things that more or less everybody is gravitating toward. The idea of a log-structured file system is much more attractive. There are many other architectural changes that we'll have to consider in disks with huge capacity and limited bandwidth.
Yep. They've been doing this for years in astronomy. Rather than running some high-bandwidth solution to some remote part of the world, and as the data isn't time critical, the main form of transfer was tape.
Also, Sneakernet scales upwards well. Want 10 terabytes? Your shipping costs go up a little, and your disk copy speed becomes more important, but otherwise, you're laughing.
This reminds me of the back-of-the envelope calculation of the "bandwidth" of medium-format (2.25" square) film. Assume film captures an image of roughly 16Megapixels in 1/1000" of a second, that's 48 GB/Second. I think that's right...
“I looked at the bandwidth bill for Wikipedia, for instance, and it is actually substantially lower in the last year than the year before, despite traffic growing by a factor of 4.”
—Jimmy Wales, http://blogs.zdnet.com/open-source/?p=899
I think you mean KB not Kb, MB not Mb, GB not Gb, and TB not Tb.
Only when talking about memory space.
Data transfer rates are still measured in bits/second -- thus, the use of a lower-case 'b'.
I'm only pointing this out because ever value, in the main post, is in bytes/second (which is inaccurate).
I can tell you this right now, sneakernet would not work anywhere in scandinavia or in any nordic european country.
Here i pay the equivalent of 24.7887 USD/month for about 1.3MB/s for my home connection. It's not even that fast with it's 10Mbit/s speed up and down when you compare it with others in Sweden. I just don't need that much at home but some have 100Mbit/s at home. Mine goes through phiber cables built into my house, most apartment houses have them here and a lot of neighborhoods with houses are working together to have them installed.
My apartment building is part of a project running all over my city. Most cities have this, even smaller ones with around 20-30000 people. It's usually a cooperation between the local electric company and the various owners of apartment buildings. These usually own large parts of the cities apartment buildings so all their houses get this advantage.
At my job we get 100Mbit/s redundant capacity from a dedicated phiber and i can't say how much we pay for it but it's about a fifth of what you have in your chart for the OC-3 with about 152Mbit/s.
I hate math but whatever the exact values are, the US people are being ripped off. :(
Perhaps the best scenario is to rent the HD from "SneakerNet provider" (delivery companies such as FedEx are in a very good position). Every time you send a box, they replace it with an empty one, ready for file copying the next delivery. The cost should be even cheaper. That way no HD travels empty. Waiting for delivery of an empty HD is like waiting for a huge download full of zeros! :-)
You write off the cost of the hard drive, but basically ignore the fact that, apart from the 56K modem, the others are the same monthly cost regardless of whether you download 1 byte or 1 TB. The cost for a 20GB file is a little meaningless, therefore, unless you assume that you saturate the connection for the entire month.
I'm paying 25 per month for an ADSL connection that can achieve 600K/48K - my ISP offers the raw connection, without some of the additional features (webspace, email addresses, static IP and better tech support), for 20/month. It's officially a maximum of 8Mbps down, 800Kbps up, but you'd have to virtually live in the phone exchange to get that (I live about 100 metres away and get a reported line speed of about 7Mbps).
I'm ignoring the cost of the phone line rental - I assume that you still want to use the line for a phone. At work we also have ADSL (running at 6Mbps reported line speed because we're a few miles from the exchange) but there isn't a phone connected, so add 10 or so on top of the ADSL cost.
I've seen 100Mbit uplinks at colo's for something like $300 / mo.
If you buy that uplink crap then you're probably also a member of the scientology church. A lot of web hosting companies tell you what uplink you get just to sucker you into ordering their package. In reality the uplink has nothing to do with what they rate your bandwidth to. The uplink is just where your cable is connected. Usually they're talking about the switch port you're connected to or the ethernet card you get in your dedicated machine.
I'm only pointing this out because ever value, in the main post, is in bytes/second (which is inaccurate).
This is intentional. I hate the "bits" nomenclature. Megabits? Really? Is this 1993? Are we using Sega's Blast Processing to transfer our data?
Bytes is the unit of measure on computers, so I'm using Bytes.
Jim Gray may have been insightful, but Tannenbaum said it a lot earlier.
GeekTieGuy - It was an eye opener for me also when I read it way back then, and funny enough for it to stick in my head to remember it so many years later.
Bengt: I have 100/10 (199 SEK), but I can get 100/100 for 249 SEK, with Adamo. I've heard others who can get 100/100 for 129 SEK or around there. The speeds that the post speak about (not T3 and OC-3) are... Well, a joke. In Sweden atleast.
So, uh, yeah... How long would it take to walk across Sweden with a backpack full of harddrives...
Mike: I'm not sure what ISP you're using, but if I download 1 byte one month and 1 TB the next, I'm going to notice a slight difference in the bills -- about $9700 CAN. That is, $10/GB past the first 30 GB.
I'm just picturing a dump truck full of hard drives with the words "The Wireless Web" printed in a sleek font on its side.
A couple of points;
Cringely has been spending his last few weeks talking about this subject. He asserted a few of "facts" I didn't know before: more than half (I recall the number as 80%, but don't quote me; I do remember the number as more than half) of Internet bandwidth is chewed up by video already; Google is buying up bandwidth for The Big Takeover; and if Google doesn't pull it off, video will kill the Internet. I think the last is how it will turn out. Physics will win. Especially since TCP is a stupid way to send such stuff. Gad. From the same folks, generally, who think XML is Just Ducky. etc.
Seymour Cray, when asked how it was that he could build computers that ran so much faster than anybody else's said: "shorter wires". And he meant it literally.
Lotus 1-2-3 began the long slippery slope of computer morphing from computing machine into consumer appliance. It's a stupid way to send such stuff.
I just had time to skim the article, but I don't see any mention of the risk of the mailed package being lost or stolen. Where I grew up, that's a high probability.
This kind of math helps me quantify just how much monetary damage I'm doing to Valve Software via their Steam network. Every time I reload my XP machine, I choose to "install from the internet," to the tune of some 7GB of downloaded data. I have the original game CDs sitting within arm's reach--I don't even have to get out of my chair to reach them--but I load from Steam anyway. And I load up on all their awful games I'll never play, including 1.5GB of Counter-Strike: Condition Zero. Yeah, I'm somewhat vindictive.
I'm up to 4 or 5 full Steam reloads at this point.
mmmm... do the hard drives cope with the heavy handling of some couriers? I would never send one with Royal Mail, but do the private couriers actually read the "Handle With Extreme Care, very fragile" stickers?
One could also decrease the copy time and increase reliability by using some type of external raid enclosure. Even better if it connected via a gigabit ethernet card. Though initial cost would increase as well as shipping cost.
Don't forget, no chance that you'll get "File is Corrupt" when your downlaod reaches 99%!
I must admit, I never really thought of the SneakerNet as a cost-effective way to deliver data. Knowing this, I may be able to cut down on some costs myself. Thanks for the insight on that.
I've seen 100Mbit uplinks at colo's for something like $300 / mo.
That's fishy, but... Cogent sells 100Mbps pipes for $1,000/month (double that for service providers). A recent Dreamhost 'blog post pegged their costs at about $30/megabit. And my employer is running a promo at one of our colo facilities, $2,000 for 100Mbps and a full rack.
I would guestimate Amazon's transfer costs at $0.015/GB or less.
I've also seen 100Mb commits in the $800/mo range, falling rather rapidly (I think they had a Christmas special at something like $650/mo). You could obviously get multiples of these for less than the OC-3 price quoted. The vendor we contacted would run this to our office or a carrier hotel.
But point taken that the bandwidth won't scale to all applications.
When your typical business egress is well in excess of 10-15Gbit/s, the billing model for big enterprise is radically different than just leasing an OC3 and paying flat-rate per byte. Such enterprise scale deployments negotiate multiple optical carriers for multiple datacenters then buy bandwidth billed by $/mbit of their sampled peak egress at 95th percentile on a month-to-month basis.
For the privilage of having pretty much uncapped bandwidth - I've no doubt that they're paying somewhere around of $40-80 per megabit/s, at 95% percentile. (varies by datacenter and region) Consider that 12Gb/s peak egress (@95th %ile) would cost a between $600K-$1m per month.
If Jim had moved his multi-terrabyte data at wirespeed possible between datacenters e.g. untweaked a GigE typically puts out around 300-500mbit/s - he could easily move the egress 95th percentile upwards - increasing the billing for the whole datacenter. When you have multiple OC192's, latency becomes the only real bandwidth cap.
Its too easy to forget that at such scale an casual drag/drop could cost $1000's... Of course telcos are love you chewing more bandwidth - their meter is *always* running.
Those who've been in webhosting know, YouTube GoogleVideo have WAY killed the bandwidth business.
Companies like Softlayer provide computers to you with 250GB of storage and 2TB of BW for $159 a month, Up that setup to a 2 TB Raid0 and cost goes to about $289 a month. Bandwidth costs are $0.25 GB over 2TB but you can buy bandwidth in 1TB blocks for much cheaper than that. I have to wonder where people are getting their information. I'm willing to bet YouTube.com doesn't pay near $0.10 a GB, I'd imagine it must be below $0.01 a GB. I mean really. Have you really done the math on how many videos are uploaded, and played a day........
I've seen 100Mbit uplinks at colo's for something like $300 / mo.
I didn't do the math exactly, but I know it came out to be a good deal cheaper than S3.
So that puts it in the $.05 - $.10 / GB range.
I'm thinking that the internet really is like a bunch of tubes. Some tubes are faster than others. The faster tubes cost more money. However, sometimes what you want is a truck full of tapes (or CD-R's or DVD-R's or hard drives).
I've seen 100Mbit uplinks at colo's for something like $300 / mo.
Yes, but what happens when you try to upload or download a terabyte of data through that colocated network connection?
If you try it, I bet you dollars to donuts you get slapped with a hefty extra bandwidth fee.
For most ISPs, "unlimited" doesn't actually mean *unlimited*. It means you are expected to use the typical amount of bandwidth, and if you exceed the allocated amount, you're metered.
i really varies from country to country
I don't know about cost, but I've used "sneakernet" several times because the performance is better. If I had 100Gib on Internet2, that would evaporate, but I'm in the US, and my company has a T1, so transferring even a few hundred GB takes a while. A T1 delivers:
~11.5 MiB/m or
So it takes a couple months to upload 1TB, whereas with Sneakernet you put it on a hard drive, and ship it in a day or two. So this make a lot of sense for US companies dealing with large datasets--unless you're one of the select few with a true next-generation connection, sneakernet is faster.
Whether it's cheaper is less clear, but it looks like it's at least competitive. The numbers that are being thrown around, like 100Mib/s for US$45, are probably:
1. Consumer prices (not available to companies).
2. Restricted by terms-of-use agreements, preventing this sort of wild and sustained bandwidth usage.
3. Download speeds, with vastly slower upload speeds. It doesn't help if you can download at 100Mib/s, if the sender can only upload at 1Mib/s.
I'm speculating on all these, and please clarify if these aren't the case, but that's how it works in the US. You can get a good price on a pretty fast download pipe (in the US, that would be 5Mib/s), but the upload speed will be much slower (often under 1Mib/s), and they will cancel your account if you do anything like full utilization. They're not intending to let you use 100Mib/s sustained; they couldn't possibly afford it at that price. They're thinking you'll a few kib/s sustained (average), with bursts of 100Mib/s, so they can cram you and 100 neighbors into the same 100Mib/s slot. To really see what bandwidth costs, you cannot look at consumer prices, you have to see what they charge businesses for connections that can be fully loaded (they'll still double those up, but not nearly as much).
When I first read about the big telescopes doing this, and wondered why they didn't use the Internet, my collegue and I ran the numbers and realized how much faster it was to ship tapes. Our conclusion: great bandwidth, but lousy latency!
Several people asked what happens if the package is lost or destroyed in transit. Well, what if the packet is lost in transit? TCP re-sends it. In sneakernet, you are playing the part of TCP, so if it's lost or destroyed, you hear that from the recipient, and you send another one. Packets are rather expensive in sneakernet, but if your network is of a quality that few packets are lost, the average cost to re-send one won't be high. There's a cost in time too, just like TCP, but where TCP can resend the packet in a few ms, it will take you a day or more. If you want to reduce the change of a catastrophic loss, follow TCP's example and use smaller packets; send it on individual 100GB hard drives, in separate shipments, and resent only those that don't arrive intact.
Great bandwidth, lousy latency!
Your DSLs are messed up.
Here in Finland atleast it's a lot different, i pay for 8/1Mbps home DSL 45€ per month, or roughly 59USD and you can get with 85€ per month you can get 24/3Mbps ADSL2, and ain't the cheapest place.
85€ comes to about 100USD per month with exchange rate of 1.3.
Also the same expensive place offers 3/1M DSL for 36e per month.
So at best, price is doubled, but bandwidth -6- times more upstream per dollar :) On top of that, you can get 2.3/2.3 SDSL for about 60€ per month if you are lucky :)
Also, there are some restricted area offerings for way better rates (think under 30€ per month for 10/10) :)
DSL aren't bandwidth capped nor restricted here at all :) and on top of that, here you can get an 10Mbps Colo place for 150€ per month.
Also like Colo Owner said you can get with softlayer for 0.10$ per GB, but best deals i've seen have been 0.03-0.06$ per GB.
Yeah, as some people have already mentioned, the prices are way out of whack for what I'm used to. I'm paying $28/month (after the extra communication fees they tack on) for 6 mbit downstream, 768 kbit upstream. Since 8 kbit = 1 KB, that means both my upload and my download are 4x what's listed for basic DSL, and the cost is 7% less. The average cost per GB, then, is about 6.6 cents each way:
Average cost each way per the original article: (0.06 + 0.51)/2 = 28.5 cents
(28.5/4)-(.07 x 28.5/4) = 6.62625 cents average cost each way.
Let's ignore the costs of the tape media for the SneakerNet. There's still another big factor to consider:
You're paying the phone bill whether you're using the connection or not. I haven't come across a connection that charges per bandwidth since I first got cable in 1996 with @home. So you're left with the choice of either A) spending $1,500/TB for the Sneakernet or B) paying that $2,000 for your 99% unused bandwidth during the time in which you could be transfering that TB of information, at no extra cost.
The only time I could imagine this being practical, barring a per-bandwidth charge on your line, is if you're transferring more than one huge set of data at a time and need it all to be done quickly.
Please feel free to point out any errors in my logic... I'm really not seeing the efficiency of a sneaker net, though.
One thing I should reiterate that a lot of people seem to be overlooking:
1MB is not the same as 1Mbps or 1Mbit. MB = megabyte, whereas Mbit, Mbps, Mb, etc = megabit. Internet service providers always list the speed in bits or bauds. There are 8 bits in one byte, so an 8Mbit/1Mbit connection is actually 1MB/128KB in the terminology of the original article. Because the topic is the transfer of terrabytes of info, the author presumably did the conversion off the bat to simplify everything.
I accounted for that discrepancy in my previous post, though.
i pay nz$99 for 40gb bandwidth on a cable plan :( New Zealand is very expensive. Over here, sneakernet would probably be worthwhile for large corporations.
Wasnt there a competition a few years ago as to who could transfer a TB of data the fastest(from london to paris iirc).
i think the winner just airmailed a box of HDDs.
and then i read about a rant for a p2p network based off a similar idea, except on a much more local scale (couriers, dead drops etc)
with flash memory getting cheaper and cheaper(and internet data transfer costs not really) i think the old mail service might be in for a revival
Hi all, long time reader, first time commenter. I've put together a calculator where you can play with the various assumptions above to see how the numbers would change.
Thought you would find this useful...
Does that mean NetFlix's business model of movies by mail will still be a good business model?
All you kids are forgetting to take into account one critical factor: pollution! If we all use SneakerNet to send things around the world, we are increasing the amount of CO2 and otherwise pollution into the atmosphere, further destroying the ozone, making the Earth warmer, making even more hippies even more upset, making Bush-Family/EvilOilCorporations even more filthy rich. Iraq war rages on (if it's for oil). Taxes go up (in the US). More people die. Skin cancer danger goes up 10,000% from no ozone leaving us to move underground where we'll just end up building superfiber connects everywhere (hopefully) like Sweden and Japan have done it (and how it should be in the U.S. if all that Telco tax-cutting that happened years ago to "build a great and all powerful nation-wide infrastructure" had actually happened instead of them trying to get even richer with this anti-Net Neutrality crap, etc.
Reminds me of the truck full of videotapes from A. Tanenbaum's "Computernetworks". The truck has a high bandwidth, but low throughput.
The pricing seems outdated. With european web hosts offering $0.10 and less per GB overage to every customer, I doubt that YouTube pays that nowadays.
With bandwidth and flash memory becoming cheaper each day, it will be interesting to see how this calculation works out in 1 or 2 years.
That's not ACCURATE AT ALL!! I have Verizon DSL and I get about 350 KB/sec Downstream and 60 KB/sec upstream (3360 Kbit/sec down and 768 KBit/sec up).. And I only pay $30 a month.
David, that's excellent pricing. I have SBC DSL, 1.5 Mbit down 384 kbit up, and pay $25. Anyway, ZMODEM, that brings back memories...and we were lucky to have ZMODEM (it was so much better than xmodem)
Just for completeness (I can't find the %-age quote), this Cringely on 19 Jan.
Looking at this problem from another angle, right now somewhat more than half of all Internet bandwidth is being used for BitTorrent traffic, which is mainly video. Yet if you surveyed your neighbors you'd find that few of them are BitTorrent users. Less than 5 percent of all Internet users are presently consuming more than 50 percent of all bandwidth.
"Never underestimate the bandwidth of a station wagon loaded with tapes barrelling down the highway at 70mph." - a quote that stuck in my mind from many years ago.
Shouldn't your cost calculation of network transfers include BOTH sides of the connection, since sender and receiver must both have high speed connections for a fast Net transfer to occur?
This would essentially double the cost of the Net transfer relative to the physical delivery option.
In australia prices for adsl/cable bandwidth costs anywhere between $3/GB to $150/GB (Tel$tra) - so the sneakernet works even better!
Let's add cell phone bandwidth to these calculations. At $0.03/KB, transferring a terrabyte would cost you roughly $300,000. At 2.4KB/sec that should take ... uhh ... My calculator can't go that high.
Your prices are completely out to lunch on the costs of various internet connections. To give you an idea, our home is paying less than $35 USD each month and receiving sustained 5Mb down and 1Mb up per second. And if I chose to I could have 10/5 for about $60.
The problem is bandwidth capping, most ISPs will charge you a fortune for bandwidth over and above your usual usage. That said, I'm also not a business user, but I can imagine that a business internet plan with similar specifics and -much- higher bandwidth shouldn't cost even a fraction of the prices you have listed.
premium dsl or cable dousn't even reach a megabyte a sec where you live?
Premium DSL does a hefty 20Megabits per second down, and 8mb up over here. (that comes down to about 1.4MB per sec true download spead)
Premium Cable scales up to 22mb at the moment, (8mb up)
This is in holland btw. (which is NOT a province of germany)
Great article, but I do have one thing to note -- the T1 and T3 prices seem to be port only, meaning there's additional local loop costs that would be added. For a T1, that can be between $200 and $500 a month, depending on one's distance from the CO. For a T3/DS3, you're looking at between $2500 and $4000 a month depending on distance from the CO. And in some cases, a DS3 requires buildout by the local phone company, which can run in the low five-figures.
On a raw bandwidth calculation it's on target, but when it comes to T1's, DS3's, and OC-3's it's the loop that is the killer these days.
ATT fiber is 29/mo for 5 meg up/down -- it is artificially capped at this rate. It will do 100! meg up/down if you get it with IP-TV. (More$) but still under $70/mo.
Nothing has bandwidth like a 18 wheeler full of DVD's
Interesting related article, "Don't expect video to exhaust fiber glut":
Cisco says that in 2010, just 20 homes using the latest broadband technology to access video content will generate enough traffic to equal the entire load on the Internet in 1995.
Juniper says YouTube already generates traffic equal to the entire Internet load in 2000.
No need to ship harddrives, they are too expansive, when you can use blueray/hd dvd instead. If even that is not enough, ship more than one. They weight less too.
I stumbled upon this blog entry before, and wanted to add a little comment to note that Cogent (which is both loved and hated, but I've found their network to be pretty solid) sells extremely well-peered bandwidth (they are a hair's breadth from tier 1 status, and have been as high as the #2 most peered ASN) at $10/Mb in 100Mb intervals.
If you manage to saturate that, you are paying a hair over 3 cents per GB.
The only time I've heard of people getting a better deal than that from an ISP was when they were 90% inbound traffic, which is obviously the opposite of a typical customer pushing web traffic out.
These numbers are very expensive! I work for Cogent Communications, and our flagship product is 100 meg for $1,000. Only $10 per meg. We provide internet access for roughly 20% of the world's internet traffic. If anyone on the East Cost of the US is interested, feel free to give me a call at (212)625-4791
That reminds me of what a college professor used to say:
Never underestimate the bandwidth of a Ford Econoline Van. It's true. Especially if you fill it up with hard drives. You can move way more data with an Econoline data in a week than you could with any network technology over months.
Doing these calculations you have to sum some 4-8 hours that would be needed to write and read 1TB of data to/from HDD (I'm assuming they would by commodity SATA drives, not 15K RPM SAS drives).
BTW. doing same with SDXC cards might be interesting alternative. 32GB SD card are already available, 64GB soon to come. They only weight 2g (0.07oz or 0.004lbs). So you could stuff 36TB (72TB with new ones) in your 5lbs package.
@Sam -- we're in 21th century, cell phones have HSPA (14Mb/s) or even HSPA+ (42Mb/s) and there are flat data rates.
Why would you ship HD-DVD/bluray disk, which can't be read nor written by most computers, while you can ship SD cards, that weight less, have more capacity and can be read and written on any modern PC?
My company (webcasting) moves terabytes of data around via sneakernet. We've got these specialize machines, basically a 3U server in a ruggedized enclosure, that we transfer A/V data with. Besides the cost savings, the ability to 2TB of data from a customer location to our offices in 2 hours via currier is fantastic. Sneakernet has also proven to be far more reliable than Internet transfers for smaller volumes of data. (Ever kick off a download/upload in the evening and find it failed for whatever reason the next morning?)
The Internet geek in me would like to see everything happen over the net, but the logistical and economic realities clearly favour sneakernet.
So what happens when UPS loses your drives or they are damaged in shipping?
I think there was an experiment similar to this where someone tested the internet speed to a carrier pigeon, and somehow the pigeon had a higher "bandwidth."