Bandwidth of a SUV full of DVDs

Details

Written by Ivey Brent Laminack Ivey Brent Laminack

Category: Technology Technology

Published: 11 March 2014 11 March 2014

First the Megabytes

A friend of mine owns a Chevrolet Suburban. The measurements were taken on his unit. (Thanks William and Josh). Folding down the back seats, the cargo space is 106 inches long, 62 inches wide and 36 inches high. DVSs seem to fit nicely in a 5" x 5" footprint. At MicroCenter they were selling spindles of 100 CDROMs in a case about 8" high. I'll assume DVDs are about the same thickness. Rounding this says that on the floor of the suburban would fit a matrix of 12 by 21 CDROMs and we could stack them 450 high if we had a suitable spindle. So the suburban could hold 12x21x450=113,400 DVDs. What is the data capacity of such? At 4.7 Gigabytes per DVD, we come out with 113,400 x 4.7 = 532,980 Gb Megabytes, 532.98 Terabytes.

Now the Seconds

Consider a one-way trip from Washington, DC to San Francisco California (MAE East to MAE west). My road atlas gives 2890 miles (about 21 miles shorter than New York City to Los Angeles). Assuming we have two drivers and haul as fast as legally possible (averaging 50 MPH) this gives us a driving time of 2890/50 = 57.8 hours. This is reasonable, because the winning time in the Fireball Baker Sea-to-Shining-Sea Memorial race is typically under 36 hours.(This is the race that has inspired such cinematic bombs as Gumball Rally and The Cannonball Run) How many seconds is this? 57.8 hours x 60 minutes/hour x 60 seconds/minutes = 208,080 seconds.

  Bandwidth = Megabytes / Second

532,980 Gigabytes / 208,080 seconds = 2.56 Gigabytes/Second.How does this compare with fiber?On October 15th, 2003, CNN reported that CalTech and CERN broke the internet speed record with a sustained transfer rate of 5.44 Gigabits/second. Allowing for the 8x between bits and bytes, it seems that the SUV has about four times the current maximum network bandwidth.This illustrates that bandwidth isn't everything, one needs low latency as well to be useful. A week of turn-around time of data going from coast to coast and back isn't very useful, no matter what its bandwidth.

Bandwidth of a Prius Full of Thumb Drives

Details

Written by Ivey Brent Laminack Ivey Brent Laminack

Category: Technology Technology

Published: 11 March 2014 11 March 2014

First the Megabytes

According to the what I found online, the Prius has a cargo capacity of 21.6 cubic feet. A mid-high level thumb drive is the Patriot 64Gb drive. Product Dimensions: 2.1 x 0.9 x 0.4 inches ; 0.3 ounces. So each thumb drive occupies about 1.1 cubic inches. Allowing for some slack between the drives and not worrying too much about the close-packing problem, that gives us a capacity of 33,931 thumb drives which will weigh 636 pounds and cost us about $1.7 million at Amazon's price of about $50 each. We could probably get a volume discount, though. The Prius has a total payload capacity of 825 pounds, so our driver will have to weigh 189 pounds or less. The total data capacity will be 64Gb X 33,931 or about 2.17 Petabytes.

Now the Seconds

Consider a one-way trip from Washington, DC to San Francisco California (MAE East to MAE west). My road atlas gives 2890 miles (about 21 miles shorter than New York City to Los Angeles). Assuming we have two drivers and haul as fast as legally possible (averaging 50 MPH) this gives us a driving time of 2890/50 = 57.8 hours. This is reasonable, because the winning time in the Cannonball Baker Sea-to-Shining-Sea Memorial Dash is typically under 36 hours.(This is the race that has inspired such cinematic bombs as Gumball Rally and The Cannonball Run) How many seconds is this? 57.8 hours x 60 minutes/hour x 60 seconds/minutes = 208,080 seconds. The trick here is the two drivers. The Prius only has extra capacity for 189 pounds after the thumb drives, so our drivers will have to switch out at fuel stops, pony-express style.

  Bandwidth = Megabytes / Second

2.17 Petabytes / 208,080 seconds = 10.44 Gigabytes/Second.How does this compare with commercial internet providers. Apparently Sony makes 2 GigaBit/second internet available in Japan. Give the 8bits/byte difference, our Prius is about 40x faster than that. This illustrates that bandwidth isn't everything, one needs low latency as well to be useful. A week of turn-around time of data going from coast to coast and back isn't very useful, no matter what its bandwidth.

A Brief History of Content Management Systems

Details

Written by Ivey Brent Laminack Ivey Brent Laminack

Category: Technology Technology

Published: 14 November 2008 14 November 2008

Back in the mid-1990s, we were trying hard just to get HTML pages to display properly. About the only dynamic pages were e-commerce sites. Everybody was trying to get e-commerce sites going with Perl, Cold Fusion or whatever else they could find. By the late 1990s, other languages better suited to the web were coming out, like PHP and people started figuring out that it was a good thing for site owners to be able to edit content on their own sites. People started writing content-management systems (CMSs). This allowed people to upload photos, write stories, and made web pages much more interesting. In those days, everyone wrote their own. This was the dawn of the custom CMS. Then some people started commercializing their CMSs and building businesses that sold and supported CMSs. Several of the web-based CMSs were outgrowths of another category of software: Document Management Systems. These kept up with word-processing documents, spreadsheets, presentations: typical desktop-software documents. These were coming to the fore around the year 2000. They were used by large magazines and newspapers, and generally ran to a six-figure implementation budget. In the early 2000s work began on open-source CMSs: Mambo, Drupal, and others. For the first few years, they were only marginally useful, but by about 2004, they were starting to be ready for prime-time. That's the history of CMSs: Custom (c 1997), Proprietary (c. 2000) and open-source (c. 2004).

I'll make the case that there is absolutely no reason to start a project on a from-scratch custom CMS these days. The last site I knew of that went down that path were left high and dry after nine months when the company decided they didn't want to support that software any more. Are proprietary CMSs worth while? It depends on the feature set needed. I'd say for over 90% of the sites out there, an open-source CMS will do the job nicely. Unfortunately, this leaves only a niche market for proprietary CMSs.

I wish I were graphical enough to draw the overlapping bell curves of:

Custom CMS  -> Proprietary CMS -> Open Source CMS

See Wikipedia's list of various CMSs