Slashdot Mirror
Mailing Disks is Faster than Uploading Data
CowboyRobot writes "Who would ever, in this time of the greatest interconnectivity in human history, go back to shipping bytes around via snail mail as a preferred means of data transfer? Jim Gray would do it, that's who. And we're not just talking about Zip disks, no sir. We're talking about shipping entire hard drives, or even complete computer systems, packed full of disks.
David Patterson (one of the developers of both RISC and RAID) interviews ACM Turing Award winner Jim Gray." Back in school we always had a saying, "Never underestimate the bandwidth of a station wagon filled with backup tapes." Seems like that still holds true.
This reminds me of how data is collected for SETI@Home:
After the data is recorded onto tapes at Arecibo, they are shipped back to the SETI@home lab in Berkeley, California. The data are then broken up into workunits, which are sent out to the client screensaver program for candidate signal detection. So far, SETI@home has generated 189,598,882 workunits from the data received from Arecibo. SETI@home has split 1,139 tapes, meaning that the average tape yields 166,709 workunits. This is somewhat lower than the optimal yield of roughly 200,000 workunits per tape because of radio frequency interference, gaps in recording, problems with the recording equipment, etc.
I think a work unit is 65,536 bytes. Even if it takes a week to ship one tape, you can't beat that throughput! But the latency is the worst.
A programmer is a machine for converting coffee into code.
"Never underestimate the bandwidth of a station wagon filled with backup tapes."
"Hurtling station wagon", "8-track tapes".
Darn you people! How the heck am I supposed to get a proper astrophysical mental image if you consistently refuse to put things in terms of multiples of VW bugs (the old ones, not the faux ones).
-theGreater
Netflix has made a business out of shipping data via snail mail, since the bandwidth isn't really there yet to do it over the internet.
Vote for Pedro
"If you're driving a station wagon around you ain't doin' too well with the ladies"
Of course playing Quake would be out of the question I would think
I wear pants.
First of all, when downloading, you have the benefit of instantly recieving the file that you need, as opposed to waiting at least a day for your shipment to arrive.
Secondly, remember that bandwidth is probably cheaper than postage. Shipping a carton with a few hard disks and proper insulation would cost at least $30 to overnight it.
Really, the title of the article comes upon the conclusion way too quickily. You must consider much bandwidth the sender and the reciever have. If both have a several gigabit OC line, then perhaps uploading it would be faster.
PING privaria.org (64.33.49.48) 56(84) bytes of data.
64 bytes from privaria.org (64.33.49.48): icmp_seq=2 ttl=242 time=2 days, 7 hrs, 37 min
64 bytes from privaria.org (64.33.49.48): icmp_seq=1 ttl=242 time=2 days, 17 hrs, 14 min
64 bytes from privaria.org (64.33.49.48): icmp_seq=3 ttl=242 time=3 days, 2 hrs, 41 min
Of course, we could put the Library of Congress holdings on it or 10,000 movies
10,000 movies? The MPAA would like to have a word with him..
Trolling is a art,
The ping on a station wagon sucks and don't even get me started on the routes...
You aren't free to do anything, until you've lost everything.
The figures, but does the cost of the bandwidth exceed the price of gas?
Eh. Guess it doesn't matter anyway. Its still cooler to be seen driving down the street w/ lots of tapes.
This is how ArsDigita University distributes its course material: http://aduni.org/drives/
This week: You can make a trade-off between latency and throughput!
Next week: Cars that can haul less can be more fuel-effiecent!
The week after: Algorithms that use more memory, but are faster to execute!
Wonders never cease!
On our Distributed Systems final, we had a question about using an airplane full of CDs being used to replace our school's internet connection. The point was the even though the plane offered 10,000 times more bandwitdh, the 80 minute latency meant it wasn't a viable replacement.
Jason
ProfQuotes
Chips have gotten faster. Ram is bigger faster and less expensive. Disk space is dirt cheap.
But the telecom industry is just crawling in comparison. I use the same phone line for dial up now as I did 10 years ago, and things like ISDN, DSL, and Cable Modems get you better performance, but nothing stellar. I don't think a T-1 has really changed in cost for a very long time.
Funny, when the bubble was expanding all the talk was about the bandwidth we were suppored to have access to, but it never made it to my house.
Eschew Obfuscation
Eschew Obfuscation
CowboyNeal writes:
That `saying' is from Andrew S. Tannenbaum's notoriously well written textbook titled simply: "Computer Networks".
It was certainly in the 2nd edition, the one I used, and might have even been in the 1st edition. I is still in the latest edition. (One of the young-uns in the office has the 4th edition on his shelf.)
A famous line if ever there was one in the geek world, although perhaps not as humourous as Chairman Bill's:
"640K ought to be enough for anyone [ paraphrased ]".
...when the modems were scarce and phone bills high. Every more or less respectable demoscene group had a member whose function was listed as "swapper".
:)
;) which said a girl wants to swap, everyone welcome etc. This was bringing a good deal of free floppies, often with some quite funny stuff on them.
:)
Swappers would get in contact with swappers from other groups, and exchange floppies full of newest stuff, productions, news, and everything of any interest (plus some exotic stuff other than floppies - a chicken bone, The Party membership ID, misprinted train tickets, and whatever interesting that caught the eye and filled the envelope up to (but not above) another price-weight treshold.)
One of the most specific swapper activities was "faking stamps". With 80 and more contacts, at least one letter a month exchanged with each of them, you had to cut on stamp prices, so you smeared the stamp with water-washable glue and wrote in the letter "stamps back", so your contact ripped your stamps off the envelope and sent you in his reply letter together with floppies. Then some washing and stamps could be reused - one set of stamps could go the same way 5-6 times before they needed to be replaced because they started looking suspect. And if it was found - you never put return address on the envelope and nobody in the post office could ever read an Amiga floppy
Another practice was making the floppies sent pretty. You almost never sent back the same floppies - they were in constant flow. Adding a marker signature was the default. Often some sticker or a drawing was common. But there were true masterpieces: A floppy painted gold, with the metal part (and under it) painted silver, the metal part without the spring but removable and attached with a thin chain to the write-protect hole, so you removed it before inserting and it was hanging from your floppy drive while the floppy was inside.
And finally all the "disk hunt" methods. Famous swappers were rarely replying to newbies who were asking for contact - you had to gain some fame on the scene with your group's productions - or get a recommendation from another swapper. So - the unanswered letters were a good supply of floppies. Sometimes they would even put an ad in some zine (spread by swapp of course
Well, Internet was what put end to it. Plus average data size - sending 6-8 floppies in one letter wasn't cheap or easy anymore, and with A1200 getting more common, high-level languages, multi-disk demos and mpeg movies, it became necessity...
Nowadays still throwing a CD across a computer lab is way faster than transferring the data over the net
45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
Basicly they shunt data around, the same way Exxon et al move oil.
Speaking as a person on a 28.8kbps connection with well over 3TB of local storage, I could only *wish* that wasn't the case.
-- I wanna decide who lives and who dies - Crow T. Robot, MST3K
If you RTA, it sounds like based on current advances, in 10 years we'll be at the point where disks are so large 920 TB each) that access will have to become sequential (making them like tape today, access speeds not increasing as fast).
That would leave room for RAM to essentially become used for random access in the way the disks are used today and perhaps current cache on the CPU to be used more like RAM is today?
A lot of wire-speed net devices are starting to look like this, with their info stored in a non-volatile storage device, but loaded into RAM on startup and all "work" done in ram.
It's easy to image a whole chain reaction of purposes for devices slipping into other functions as a result of varying levels of technological advancement in them.
The party of stupid and the party of evil get together and do something both stupid and evil, then call it bipartisan.
i believe your attribution is correct.
Never underestimate the bandwidth of a station wagon filled with backup tapes
however, while the immediate bandwidth of a station wagon filled with tapes may be enormous, the overall bandwidth is quite poor. this is because of the slow write/read rates of the tape drive, and the slow overall speed of the station wagon. i can transfer 3 gigs from my work computer to my home machine faster than the time it would take me to write the 3 gigs to tape, drive it there, and read it back from tape (and my drive is only 15 minutes). if i lived 5000 miles away, my tape bandwidth would be considerably worse, while my internet bandwidth would be virtually unchanged.
since this statement was made, we have reached the point where internet bandwidth has exceeded the "vehicle full of tapes". now, this one might be good for a few more years:
Never underestimate the bandwidth of an sr71 full of netapps
What a great example you picked! Cable TV companies are pumping dozens of digital movies accross their system at once, live. Yet they crimp your upload speed to DSL rates or lower, 30KB/s, because they are afraid of people "stealing" movies. This is not a technological problem, it a social one. Big publishers and telcos are afraid of competition and are doing everything in their power to keep you from enjoying technology that's already in place. It's the same old fight Ma Bell used to wage back when they would not alow you to so much as plug a modem into your phoneline.
How long are people here in the US going to put up with this monkey business?
Friends don't help friends install M$ junk.
uucp to Australia used to be done by uploading a spool dir somewhere in the US to a tape and airfraighting it to Oz, then doing the same at the other end. You'd post something to usenet and get a reply 2 weeks later
Not only does a stationwagon full of harddrives have a respectable sustained throughput rate, the contents don't get screened by the firewall. Ditto for the hardrives in a briefcase, or those USB drives on a keychain.
Exploding capacities of storage drives have implications on attempts to keep data within boundaries, as well as attempts to getting it from point A to point B.
traceroute privaria.org
1 privaria.org-package.ready 50000ms
2 Picked-up-USPS 900000ms
3 Transfer-to-USPS-depot 300000000ms
4 (unknown)
5 (unknown)
6 (unknown)
Packet Loss 100%
Blockwars: multiplayer and it's free.
"They do not preach that their god will rouse them, a little before the Nuts work loose." Kipling, 'The Sons of Martha'
One of the major thoughts in the whole interview is that our storage has increased to such a point that we can't access it all in a reasonable fashion. For my uses (which are far from industrial level) I find that I can only watch one movie or listen to one song at a time. On my 200 gigs of hard disk I've got 60 gigs of music (and growing daily) and at least 100 gigs of movies.
Don't judge me on the legalities of the situation, but note that this isn't uncommon...I have some very drastic media needs and the media that I like is pretty intensive, but I don't very often need to stream any of it en masse to another location. It suits it's purposes fine exactly where it is, and I haven't had any problem acquiring any of it or accessing it.
I suppose my rambled-to point is that for my needs I'd rather there was more storage at this point than have higher access speeds as I can get all that I need as fast as I need it. Perhaps our usage of the medium dictates how it develops.
"Share your knowledge. It's a way to achieve immortality." -- Dalai Lama
I just recently moved halfway across the US from my hometown. A buddy of mine who had a ton of MP3s (mostly legal BTW) had just suffered a HDD crash and his SO's car had been broken into meaning that TONS of music had been lost/toasted. Before I left, I'd copied his whole collection to my drive. Shipping him a drive with the whole contents (60 GB) of my music collection took a hell of a lot less time than letting him download it (at 20 Kb per second (Ghod I hate SBC!)) or worse yet, take the time to pick through it at human speeds, and was far cheaper unless you figure that the cost incured by me sending it overnight was in addition to my regular bills.
Dok
"You can't screw the system, but you can give it a good fondling." -- Too lazy to look it up
First of all, how could you drive into a garage at 60 miles an hour, stop, unload, and get back out before the next truck, only four inches behind you comes in? In your system the trucks would crash into eachother, and you'd get no effective bandwidth.
In actuality, you need to figure how long it takes to unload all the tapes from the truck at least. Assuming about 10 seconds a tape (ones in the back take longer) on average and that's 1.16gb/second.
autopr0n is like, down and stuff.
Is that why you posted the exact same joke as someone else 4 minutes late?
autopr0n is like, down and stuff.
But they could also take 50 different routs to get there. There are all different kinds of ways to rejigger the figures, but were talking about what's practically possible. Employing enough people to man those 40,678 loading docs full time (what you would need to offload 1 truck/ 0.28 seconds), would be at least 5.55*40,672*24 is about $1,083,667 dollars per day, or almost $400 million a year. For that kind of money you could probably afford to lay down multiple parallel multifrequency optical cables.
autopr0n is like, down and stuff.
Q:A man with a delivery bike can pedal at 20mph between the organisations two offices that are 5 miles apart. The basket on the bike can carry five half-inch tape reels. What is the effective throughput of this datalink? For extra credits: A modem can transfer data at 300bps. At what distance does this outperform person with bike.
Engineering is the art of compromise.
The station wagon comment reminded me of an idea that I had a long time ago, when I first read about how the Internet routes packets around. You know how you can ship stuff UPS overnight? It can get pretty expensive, depending on how big and heavy the package is. And sometimes, businesses would pay an even greater price to have a package delivered even faster. Why not introduce a system for getting things delivered extremely fast, and I do mean fast, all around the world?
Imagine this: Put together a network of railroad-like tracks that are enclosed in concrete tunnels. In a vacuum. Individual cars would travel on these tracks at greater than mach speeds. They would essentially go from one switching station to another, kind of like the telephone network or the Internet. They might come in several sizes, these cars. When you need something delivered fast from, say Los Angeles to New York, the package would be placed on a dedicated car which would take it at blazing speeds through, say, Albuquerque, Oklahoma City and Louisville, to New York. At each station, equipment would adjust switch tracks to route the car to its next switching station; the car would not even have to stop or slow down. The package might be there in four hours, counting the time it takes to bring the package to a station, have it loaded, unloaded, and then transporting it to its final destination.
This might actually make shipping cheaper rather than more expensive. Automatic equipment sorts mail at the USPS. If this mail were collected, say, once every hour (during business hours), taken to the nearest major USPS distribution center, where it is sorted, placed in boxes heading to the same destinations, and then shipped (tunneled?) through the above method, mail going to a distant location might arrive faster than mail going across town. This could be done with collections of packages that are all going from one major city to another together. Load them in a container and bust them all over there. Sure, it'll still take, say, 24 hours to ship packaged in such groups, to save money, since you have to wait for enough packages, sort them, group them, etc., but if you want something shipped right friggin now, the option to get a dedicated car is still available. This might reduce use of gasoline and use of air and ground traffic. If computers can control the cars on these tracks so that cars are going mach 2 almost bumper to bumper, that would allow for extremely great throughput.
Back to the station wagon comment, supposing this could be done, (running more tracks all over the world and installing these switching stations at each major city), you could load hundreds of terabytes of data onto a big friggin raid system and then get that data across the world faster than shit going through a tin horn.
I work in a AmLaw top 100 law firm in DC. We do a lot of complex litigation work. We use software such as Concordance, Ringtail, and Litgator's Notebook (runs on Lotus Notes) to manage collections of documents. The documents are scanned to group IV tiff; the meta data and OCR text that is extracted from the documents at scan time is loaded into another database that overlays the images.
These tiff file collections run into the millions.
Of course the point of doing this is to facilitate collaboration on document review between us, our clients and our co-counsel. These people are often 1000s of miles apart, and nearly as often have crap for IT resources (equipment and personnel).There are ways of accessing this stuff over the internet securely but it's never quite the same as having the real version of the software. This form of access often proves to be impractical for the lawyers who travel alot depending on the type of access they can get wherever they end up.
So what often happens is, we end up dumping the entire collection on a laptop with a big hard drive or a bigger firewire or USB drive, so they can work without access to the internet and then replicate changes when they can get the laptop back on ethernet or a POTS line.
Collections of images and databases (not to mention the various Power Point presentations and word processing files) can very easily run over 50GB. Moving this across the LAN, over my PC BUS to another hard drive and then FEDEXing it is certainly faster than doing the same transfer using FTP or SCP. Not to mention, that way I can install the software (properly) and test the whole setup before I send it off. The extra wear and tear I save on my psyche from NOT having to explain how to install all of the software, point it to the image collections, and deal with equipment I have no control over while being screamed at by extreme Type A attorneys going to trial makes that laptop look like a pretty good investment.
These are good if you have someone on the other end of your FEDEX run who know how to open the case on a PC and install a HD themselves. I can setup one machine with everything, image the hard drive, make copies on other drives and drop them into FEDEX pouches as fast as I can make 'em. I can't think of a faster way to move a few 100 GBs of stuff to a half dozen places inside of a day. If someone has ideas, I'm all ears.
If you never make mistakes, it's probably because you're not doing anything.
At some point you have to actually get the tapes into the computers. Even if the tape drives themselves had relatively infinite bandwidth, it would still take at least 10 seconds to get it off the truck and loaded into the reader. Maybe you could save time by using some sort of SUPER GIANT SPOOL like 2 meters in diameter and height.
Imagine that, cassettes the size of shipping containers. Of course, if we're going to talk about things like that, we really need to talk about tape read speed too. Lets say the tape thickness is 8 microns. In that case, each layer can hold (4-(8*10^-6)*i)*pi where i is the layer number, So the total number of layers is Sum(2*(2-(8*10^-6)*i)*pi) (2*pi*r where r is 2m - 8um*i) for i from 0 to 2/8*10^-6. that gives us about 4*10^11 meters of tape. Even if we spun the tape at the speed of light it would still take 20 minutes to read one tape. At the speed of light, it would take about 1309 seconds to load the data into the computer. Since 1309/100 is 13.9, assuming you did your math right the cost would be $278 million per year. Of course, we can't actualy spin the tapes at the speed of light.
autopr0n is like, down and stuff.
So how do they do this? I've always been under the impression that with digital cable and cable internet, all of the data has to be sent to everyone (in the same neighborhood anyway), so how can they handle the hundereds of channels (some of which are actually lower quality than others), the multiple VOD streams (even for the same movie), and eveyone's porn and mp3 activities all at the same time?
This one is simple they ran fiber to the curb a few years ago. They even ran new coax to our apartments to handle more bandwidth. There is effectively infinite bandwidth running into your apartment.
"Today disk-capacity growth continues at this blistering rate, maybe a little slower."
What is a bit slower than a blistering rate? A skin-reddening, sensitizing-to-the-touch rate?
RTFM; please, I beg you.
Favorite quote from the article: "Not many of us know what to do with 1,000 20-terabyte drives--yet, that is what we have to design for in the next five to ten years."
Heh. I do, so get designing. The various law firms reviewing documents from cases like Enron (criminal , bankruptcy, and civil procedings), Microsoft's antitrust suit, the SCO v. IBM, etc. etc. need that space to store all the materials from their case work. Lots of paper from all those places get turned into electronic images managed by very large custom databases.
Guess how many Group IV tiffs and pdfs some of these become. Answer: millions. In five or ten years, cases such as these will likely consist of collections of data that large. Terabytes of data for cases such as these are not uncommon now. Enron could get this big by itself by then. It's well on its way to becomming one of the largest cases of all time. Check this out. Whoa.
If you never make mistakes, it's probably because you're not doing anything.
Never underestimate the bandwidth of a station wagon filled with backup tapes.
;-)
This is a statement by Andrew S. Tanenbaum from his book titled Computer Networks. Though it's supposed to be a text book (with 4.5 stars on Amazon.com), I and most of my friends also regard it as a nice collection of stories related to computer networks and communication
I'm pretty disappointed (although not entirely surprised) in SlashDot posters. This article was clearly more than simply 'mailing disks' which > 95% of the topics (including dupes of dupes of dupes of ...) on this article have been about.
/. audience.
Sure, he mentioned cost of shipping disks, and actually concluded that shipping an entire computer system is more economical than mailing individual disks. However, there are far more interesting and discussion-worthy conclusions he raises.
What about disk capacity reaching such incredible sizes as 2TB/disk - and the fact that current random-access methods will render such drives unusable? This affects all of us, since our OS' filesystems will need to fundamentally change to be more sequential (e.g. like tape drives). Personally, I hope that whatever happens to the fs the OS will insulate me from being forced to use it in a sequential manner (e.g. will I be exposed to the sequential nature of the medium or can it be successfully abstracted?)
He talks about, in almost glowing terms, the SlashDot favorite MySQL and how "At some point, somebody will say, 'I'm running my company on MySQL.' Indeed, I wish I could hear Scott McNealy [CEO of Sun Microsystems] tell that to Larry Ellison [CEO of Oracle]." And, although the Research Area people are pretty independent, this is from a MICROSOFT employee. Not a peep from the
Personally, I think that using MySQL as a 'research tool' as he suggests is a Very Bad Idea - it's not even a mediocre implementation of the relational model and there are better open-source implementations out there (PostgreSQL being the one that comes to mind). Basing scholarly studies on MySQL would be like basing the foundation of a skyscraper on a shack (not that any other SQL DBMS's are much better, but why use one of the worst?). The best 'research vehicle' would be an open-source truly relational database management system (there are no commercial TRDBMS either). It doesn't have to be very advanced, but it has to be architected from the ground up to be a TRDBMS (which means SQL doesn't cut it as a query language).
One thing he notes which I see as being a large problem in the open-source community as well is how "...The thing that slows Oracle, IBM, and Microsoft down is the testing, and making sure they don't break anything--supporting the legacy. I don't know if the MySQL community has the same focus on that." As a long-time PHP developer and advocate I'm still hesitant about updating our production systems - it seems as if every successive release of PHP has innumerable functions removed or changed with no ability for backwards compatibility. I guess it's a lot easier to say to users 'you get what you pay for' when they are just that - users and not clients. One of my disappointments from many open-source proponents (which I am one) is the hostility to treating clients as clients - 'you can always edit the source', etc. - for the most part large companies don't care/want to edit the source - that is what they want to pay you to do. Until more projects (MySQL included) start to realize this, then they will pretty much always occupy niche roles in the enterprise.
Finally, even he, an academic seems to (at times) confuse the relational model's implementations' details (e.g. the SQL product performance) with the model itself (of which there is no mention of performance, because it has nothing to do with the model). Theoretically, a TRDBMS should be faster than the SQL implementations we have today. It just takes someone to do it, and I don't see why the open-source community can't build the BEST mousetrap there is - we just have to abandon the 'mob culture' of MySQL.
Thanks,
--
Matt