Slashdot Mirror
New Fundamental Law of Network Economics
intersys writes "A new fundamental law of economics has been formulated by Rod Beckstrom, former Director of the National Cyber Security Center. In Words: The value of a network equals the net value added to each user's transactions (PDF) conducted through that network, valued from the perspective of each user, and summed for all. It answers the decades-old question of 'how valuable is a network.' It is granular and transactions-based, and can be used to value any network: social, electronic, support groups, and even the Internet as a whole. This new model or law values the network by looking from the edge of the network at all of the transactions conducted and the value added to each. One way to contemplate the value the network adds to each transaction is to imagine the network being shut off and what the additional transactions' costs or loss would be. Beckstrom's Law replaces Metcalfe's law, Reed's law, and other concepts which proposed that the value of a network was based purely on the size of the network (and in the case of Metcalfe's law, one other variable)."
As whiskers abound,
The network is found,
Infinitely sound,
Until Slashdot-ground.
Burma Shave
Get thee glass eyes, and, like a scurvy politician, seem to see things thou dost not.--King Lear
1(4)+1(2)=6
Me too. I skimmed through the paper; he just defines some functions, and he calls his new model Beckstrom's Law. How is his definition a law?
a new law that will, when the current crisis will end, contribute to create the next crisis?
He's with the National Cyber Security Center, and for security purposes there's always this dramatic "the hacker caused damages of X dollars" where X is very large. They want X to be huge so that equally large sums can be spent on bringing offenders to justice, so that the press has this huge amount of loss to report for dramatic purposes, so that huge civil suits can be brought, etc. Additionally, this Beckstrom fellow devises "Beckstrom's Law" and now he can be called in for expert testimony, he being the expert because "Beckstrom's Law" is named after him (by himself of course, but that's just a minor point).
Run and catch, run and catch, the lamb is caught in the blackberry patch.
Also known as the Law of 24:
Networks are easy, simply "open a socket". (And have a gun. And Jack needs to yell at the network a lot)
Jack: "WHY ARE YOU NOT ROUTING THESE PACKETS! "JUST TELL ME WHERE THE BIT-BUCKET IS!" "WHO ORDERED THE 3COM HUB?"
What? It makes as much sense as the article.
Sent from your iPad.
I think this definition is pretty damn useless -- how is one supposed to calculate this value for anything but trivial example cases? You would have to determine the value of each transaction, and then the 'value-add' of the network for that transaction, as determined by the user. I make 'transactions' (financial and otherwise) on the Internet all the time, and I couldn't begin to guess at useful values for these. And I'm just one of millions of such users.
Finally, how would one even begin to define 'value' for the transactions in a social network? How much (or little!) is being poked worth?!?
By naming it. I have this definition of contributions to /. that increases your Karma. I call it Opportunist's law.
See? It's easy!
No, I didn't prove it in any way. That's not part of the exercise. Naming it is. So make something up and call it azgard's law. You'll see, it's fun, and you'll be the hit at parties when you can say that you're that guy the law was named after.
You should omit that you did the naming.
We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
But surely there are replacement wives available. So is the value of my wife in the positive things that I get from her currently, or is it only relative to the potential value of other wives, or of the freedom of having no wife at all?
Clearly, sir, you need a Redundant Array of Inexpensive Wives (RAIW). Twins (RAIW-1) are OK but RAIW-5 with lots of hot spares works better. Most admins agree that products with large rack mounts are better. A lot of hot air is generated resulting in increased cooling requirements. Also an astounding amount of noise, OSHA requires earplugs in that environment. This solution is popular in certain datacenters in Utah. Some folks claim a competing product exists, the Redundant Array of Inexpensive Girlfriends (RAIG) but everyone I've met agrees it usually ends up pretty expensive, the opposite of the original acronym, and there are often serious interoperability and EM compatibility issues. Finally w/ regards to financing there is considerable debate about rent vs purchase, short term lease vs long term lease, mileage reimbursement, etc. Rent to own agreements usually don't work out. Maintenance costs are somewhat beyond the realm of this email, but can be extraordinarily high. Anyway good luck with your network, Sir.
"Science flies us to the moon. Religion flies us into buildings." - Victor Stenger
It's handy that he named his observation "Beckam's Law" for us, because then we can apply inviolet's law:
A new discovery is headline-grabbing FUD/bunk if its discoverer names it after himself. Whereas a new discovery is probably useful if others name it after its discoverer.
Think of it as an extrapolation of the importance of peer review. Pretty handy, eh? That's why I named it after mys-- aw, damnit!
FATMOUSE + YOU = FATMOUSE
And surely Beckstrom must've realised this, since its trivial to get to Metcalfe's law from his equations.
Beckstrom:
Vj = Sum(i=1..n, V[i,j]) = Sum(i=1..n, Sum(k=1..n, B(i,k)) - Sum(l=1..n, C(i, l)))
= Sum(i=1..n,Sum(k=1..n, B(i, k) - C(i, k))).
= Sum(i=1..n,Sum(k=1..n, Sum(z=1..n, B(i, k) - C(i, k))/n ))
Let A(i) = Sum(z=1..n, B(i, k) - C(i, k))/n , the average benefit to user i of the network. Then:
Vj = Sum(i=1..n, Sum(z=1..n, Ai))
= Sum(i=1..n, n*Ai)
= n* Sum(i=1..n, Ai)
I'm sure you can see where I'm going with this now...
Let A = Sum(i=1..n, A(i))/n , the average benefit to any user of the network.
Vj = n^2 * A. Oh wait - that's Metcalfe's law. All Beckstrom's done here is give an expansion of the average benefit per user.
Wait, wait, I have a similar epiphany in another field!
"The value of a pizza equals the net value added by each topping, multiplied by the cheese, rated by the taste experience of each user, and minus 100 if it came from Domino's."
There you go - please send accolades and research grants this way.
On one hand, you have to put it in context. There have been several models of network value, and none of them have taken this "common sense" approach of defining value in terms of what you get out of the network.
But before running with that, we have to look at the other hand, and see why none of the older models take this approach. I think it comes down to the question of "what is the model's purpose"? The simplicity and usability of a model are potentially as important as its accuracy.
If I can't predict the inputs to my model, then my model can't be used for prediction. I can predict a certain level of information about a hypothetical network using size-based models.
Moreover, if I can't measure the inputs to my model, then I can't use it at all. Claims that we can "measure the value of the Internet" are a bit much.
Don't get me wrong, this model may be well-suited to some purposes. But it does not, as TFS claims, "replace" the existing models any more than relativity "replaces" classical physics. Sure, it may be a highly accurate definition of the value of a network when you can observe that network empirically (but a definition is not a "law"). This assumes that it accounts for negative value of a transaction, though. (What is the contribution of spam to the value of the Internet?)