Mailing List Archive: NANOG: users

CDNs should pay eyeball networks, too.

1 2

View All

Index | Next | Previous | View Threaded

Valdis.Kletnieks at vt

May 1, 2012, 2:49 PM

Post #26 of 36 (282 views)
Permalink

Re: CDNs should pay eyeball networks, too. [In reply to]

On Tue, 01 May 2012 14:27:50 -0700, Mike Hale said:
> > Mike - please get mail software that does correct quoting. It's 2012, and
> > proper quoting has been understood since the mid 80s. There's *really* no
> > excuse for using software that can't get quoting and citing right.
> *eye roll*
> Really? You wasted 36 words on this?

Wasn't wasted - this time your message included > in the right places. :)

And yes, proper attribution *is* important.

dmiller at tiggee

May 1, 2012, 3:03 PM

Post #27 of 36 (282 views)
Permalink

Re: CDNs should pay eyeball networks, too. [In reply to]

On 5/1/2012 5:20 PM, Valdis.Kletnieks [at] vt wrote:
> On Tue, 01 May 2012 14:13:01 -0700, Mike Hale said:
>
>>> "But you *may not* tie your
>>> price to the hours used to produce it for the first."
> The above was William Herrin's comment (quoting level fixed by me).
>
> Mike - please get mail software that does correct quoting. It's 2012, and
> proper quoting has been understood since the mid 80s. There's *really* no
> excuse for using software that can't get quoting and citing right.
>
>> Sure you can. How else do you determine what the software's going to
>> cost if you're not going to factor in development?
> You missed the point - having given customer #1 an invoice that included
> a line item for 1,432 hours of R&D at $221/hour, you're treading on thin
> ice if you present another customer an invoice that includes a line item for
> the same 1,432 hours of R&D (absent an agreement between the two
> customers to share the costs, etc).
>
> And if you've *collected* that $316,472 from the one customer, it's somewhere
> between sleazy and skanky to include that $316K in the costs that need to be
> amortized over the next N sales of the software.
>

From an accounting perspective, every R&D effort that I have seen or
been a part of was not billed to any customer. R&D has always, in my
experience, been an internal charge against a company's own profits.

As to hourly software development, I have never seen or been a part of a
software development effort where the final work product was not owned
by the entity paying for the hourly development. The contract software
developer can't sell the same software twice because they don't own it
to sell it twice.

It is possible that my software development experience, while broad, has
missed a model where R&D is billed to customers or software isn't owned
by those who paid to have it developed...

If a company expends their own resources to develop a piece of software
(i.e. a product) regardless of whether through R&D (internal
development) or contracting with someone else to develop the product
(external development), then they can set the price of that product to
whatever they like. There is nothing sleazy or skanky if the price of
the product multiplied by the number of copies sold is positive after
subtracting the price of the development of the product - we call that
profit.

-DMM

Valdis.Kletnieks at vt

May 1, 2012, 3:18 PM

Post #28 of 36 (288 views)
Permalink

Re: CDNs should pay eyeball networks, too. [In reply to]

On Tue, 01 May 2012 18:03:06 -0400, David Miller said:

> From an accounting perspective, every R&D effort that I have seen or
> been a part of was not billed to any customer. R&D has always, in my
> experience, been an internal charge against a company's own profits.

RIght - and when pricing the result of the R&D, you take into account the
internal charges and estimated sales volume to determine a target price
point. R&D was $316K and you estimate you can sell 100 of them, you're
going to have to charge at least $3,160 a copy to make a profit on the
project. You sank $150M into R&D and think you'll sell a million units, you'll
need to charge $150 to break even. And so on.

The point is that if you already got *paid* the $316K, it's morally wrong to
include it in the calculation of "sunk costs we need to recoup to turn a profit".

bill at herrin

May 1, 2012, 3:21 PM

Post #29 of 36 (283 views)
Permalink

Re: CDNs should pay eyeball networks, too. [In reply to]

On Tue, May 1, 2012 at 6:03 PM, David Miller <dmiller [at] tiggee> wrote:
> From an accounting perspective, every R&D effort that I have seen or
> been a part of was not billed to any customer. �R&D has always, in my
> experience, been an internal charge against a company's own profits.

Hi David,

That's called "internal" research and development or "IR&D".

It's distinguished from research and prototyping as an exploratory
effort under contract to a customer. For example, all grants under
organizations like NSF, ONR or DARPA the latter type of R&D. They want
to know if something is possible so they pay knowledge domain experts
to find out. Or more commonly, the experts submit a proposal that
says, "We think this is possible. It would be very good for you if it
is. Won't you please pay us to find out?"

> As to hourly software development, I have never seen or been a part of a
> software development effort where the final work product was not owned
> by the entity paying for the hourly development. �The contract software
> developer can't sell the same software twice because they don't own it
> to sell it twice.

That depends on the contract. By law, the folks who wrote the software
(or the company who paid their salaries) own it. By law, contract !=
salary. Some contracts specify that the copyrights and other IP will
be signed over upon completion. With most of the contracts I've
worked, the customer doesn't get the copyright, but I make no claim
that's typical.

Regards,
Bill Herrin

--
William D. Herrin ................ herrin [at] dirtside� bill [at] herrin
3005 Crane Dr. ...................... Web: <http://bill.herrin.us/>
Falls Church, VA 22042-3004

alex at corp

May 1, 2012, 3:46 PM

Post #30 of 36 (282 views)
Permalink

Re: CDNs should pay eyeball networks, too. [In reply to]

I can't agree with this. You are assuming a cost-plus model. Many things are market-priced.

If you are the only game in town, and you have a great product, you sell it for the most you can. You aren't a charity.

The customer always has the option to not buy your product.

----- Original Message -----
From: Valdis.Kletnieks [at] vt <Valdis.Kletnieks [at] vt>
To: David Miller <dmiller [at] tiggee>
Cc: nanog [at] nanog <nanog [at] nanog>
Sent: Tue May 01 18:18:48 2012
Subject: Re: CDNs should pay eyeball networks, too.

On Tue, 01 May 2012 18:03:06 -0400, David Miller said:

> From an accounting perspective, every R&D effort that I have seen or
> been a part of was not billed to any customer. R&D has always, in my
> experience, been an internal charge against a company's own profits.

RIght - and when pricing the result of the R&D, you take into account the
internal charges and estimated sales volume to determine a target price
point. R&D was $316K and you estimate you can sell 100 of them, you're
going to have to charge at least $3,160 a copy to make a profit on the
project. You sank $150M into R&D and think you'll sell a million units, you'll
need to charge $150 to break even. And so on.

The point is that if you already got *paid* the $316K, it's morally wrong to
include it in the calculation of "sunk costs we need to recoup to turn a profit".

mysidia at gmail

May 1, 2012, 5:06 PM

Post #31 of 36 (280 views)
Permalink

Re: CDNs should pay eyeball networks, too. [In reply to]

On 5/1/12, Patrick W. Gilmore <patrick [at] ianai> wrote:
> I love the fact Dominik says "from a CDN", then leaves Limelight's name in
> the text. :)
[snip]

So a CDN made the mistake of attempting to monetize an existing
peering arrangement without first having a signed peering arrangement
in place for the existing peering with a Mutual NDA prohibiting any
dissemination details of any of their communications/ future
negotiations/etc of peering arrangement or traffic exchanged to
unrelated third parties??

The lack of a NDA that would prohibit identifying a CDN and their
e-mail notification of depeering, whether their name was mentioned or
not, sounds quite un-Tier1 like to me.

--
-JH

Valdis.Kletnieks at vt

May 1, 2012, 5:08 PM

Post #32 of 36 (273 views)
Permalink

Re: CDNs should pay eyeball networks, too. [In reply to]

On Tue, 01 May 2012 18:46:56 -0400, Alex Rubenstein said:
> If you are the only game in town, and you have a great product, you sell it for the most you can.

Pay attention. What I said:

> going to have to charge at least $3,160 a copy to make a profit on the
> project.

*at least*. You can charge $4K, or $40K, if you think you can get away with
it. Or you can sell it for $2K, if you think you can sell 200 cheaper copies
rather than 100 expensive ones. But unless your sales price times the sales
volume is more than your sunk R&D cost, you're going to lose money on it.

(And yes, there's counter-examples. Sony did the PS/3 as a loss leader. Which
makes my point - they *knew* that sucker was going to lose money because there
was no price point at which the expected sales times the price was more than
what they spent in development. Only reason they did it anyhow was because
they had a business plan to make the money elsewhere...)

nanog-poster at axu

May 1, 2012, 10:06 PM

Post #33 of 36 (268 views)
Permalink

Re: CDNs should pay eyeball networks, too. [In reply to]

Morning,

I have no idea what's really going on at LLNW, but I thought I'd still
share an alternative view on this matter:

My understanding is that LLNW is spending tons of money to upgrade some
of their IXP connections to 100GbE in Europe. With that in mind, I'm not
that surprised if they wish to get some new income to cover those costs.
While content is king, eye balls are kings too. Go figure.

--
Aleksi Suhonen / Axu TM Oy
Internetworking Consulting
Cellular: +358 45 670 2048
World Wide Web: www.axu.tm

thomas.mangin at exa-networks

May 2, 2012, 1:58 AM

Post #34 of 36 (268 views)
Permalink

Re: CDNs should pay eyeball networks, too. [In reply to]

I (in the UK) had the same letter from LLNW yesterday, word for word.

When the peering was established, I had transit providers with strict peering policy (TATA/L3), now I have two with more open policy (HE/KPN). I assume LLNW now sees me via what is for them a peer, and see no financial reason to keep a direct session up.

However I must say that the wording of their letter is appalling. Even if they gave me 30 days notice to change my transit arrangement and did not terminate the session without warning, the tone of this mail is simply wrong. I am pretty sure my transit providers are seeing them via the same exchanges I do, so the traffic did, most likely, not even shift from interface. We did not have any issues of capacity and/or outage, so it is not that this change will save them much in opex costs neither. My peering ports are the same size as my transit ports, so they have gained anything in performance by shifting the traffic (and as I do not congest, did not loose anything neither though)

What it tells me is that they do not care about my business and prefer to force me to pay to reach their network (more than I was previously) via transit ... or pay more but less than transit using their "generous" pay peering offer .. I did not bother asking them what the cost was, my answer is NO. I will prefer to pay my transit provider, at least the extra capacity can be put to other use.

If ever I change back my transit provider to one they do not have favourable agreement with, I will think twice about peering again with them, or I may ask them for some pay peering to reflect their saving (no, I would not I am not that kind of scumbag).

As my traffic volume is clearly noise for them, I am sure they do not care at all. However, large rivers are all made of small streams, and all trees starts as seeds ( I am feeling zen this morning ... :D )

Thomas

I am glad they are spending ton of money to upgrade their infrastructure.. but so am I.

On 2 May 2012, at 06:06, Aleksi Suhonen wrote:

> Morning,
>
> I have no idea what's really going on at LLNW, but I thought I'd still share an alternative view on this matter:
>
> My understanding is that LLNW is spending tons of money to upgrade some of their IXP connections to 100GbE in Europe. With that in mind, I'm not that surprised if they wish to get some new income to cover those costs. While content is king, eye balls are kings too. Go figure.
>
> --
> Aleksi Suhonen / Axu TM Oy
> Internetworking Consulting
> Cellular: +358 45 670 2048
> World Wide Web: www.axu.tm
>

leigh.porter at ukbroadband

May 2, 2012, 2:32 AM

Post #35 of 36 (266 views)
Permalink

RE: CDNs should pay eyeball networks, too. [In reply to]

> I (in the UK) had the same letter from LLNW yesterday, word for word.

Me too.

> However I must say that the wording of their letter is appalling

Agreed.

> I am glad they are spending ton of money to upgrade their
> infrastructure.. but so am I.

Slightly odd though that they are upgrading their network and then de-peering everybody who takes < 1Gb/s from them.
I don't quite understand why a content DELIVERY network would want to do this.

I'm not sure who's content they deliver but this does not seem like a particularly great way to go about delivering it.

There was a network who commented earlier in the thread that they do 600Mb/s with them, that's not an insignificant level of traffic really, especially coming from a single CDN. I wonder if this not some slightly mis-informed exec at LLNW who thought they found a great way to extract more money to deliver content that they have already been paid to deliver.

--
Leigh

______________________________________________________________________
This email has been scanned by the Symantec Email Security.cloud service.
For more information please visit http://www.symanteccloud.com
______________________________________________________________________

patrick at ianai

May 2, 2012, 4:09 AM

Post #36 of 36 (266 views)
Permalink

Re: CDNs should pay eyeball networks, too. [In reply to]

On May 2, 2012, at 1:06, Aleksi Suhonen <nanog-poster [at] axu> wrote:

> I have no idea what's really going on at LLNW, but I thought I'd still share an alternative view on this matter:
>
> My understanding is that LLNW is spending tons of money to upgrade some of their IXP connections to 100GbE in Europe. With that in mind, I'm not that surprised if they wish to get some new income to cover those costs. While content is king, eye balls are kings too. Go figure.

Lots of networks upgrade their infrastructure. It means they are doing more traffic, which hopefully means their business is doing well. Very few - in fact, I can't think of a single network - start asking for paid peering just because they are upgrading their ports. Networks either ask for paid peering, or don't, irrespective of their IX upgrade schedule. It is based on whether they think they have power over their peers. I guess we know how LLNW feels now.

The interesting thing to me is the reversal from previous years. Most content providers have issues with eyeball networks saying "pay me for bits". Content networks have historically claimed this is silly of the eyeball networks - including LLNW. Eyeballs get paid by their customers (DSL, cable, whatever) to "reach the Internet". Content networks pay to bring the content right to the eyeball's door. Or so the theory goes.

This move belies that argument LLNW has made themselves in the past. It is not about "your customer pays you, my customer pays me." It is about who can force whom to pay (or not, as most people who have spoken up said they would not pay).

End of day, this doesn't change the way of the world. LLNW is a big network, but compared to the whole Internet, they are relatively small. There will be corner cases like this, and the market will decide who wins & who loses. <insert comment about "except those with monopoly power" or some such>

--
TTFN,
patrick

1 2

View All

Index | Next | Previous | View Threaded

Interested in having your list archived? Contact Gossamer Threads