Re: why strict Net neutrality works best: simple beats complex

Arun,

We seem to be talking at cross purposes - the QoS/Policing field is already
used in all new routers to block spammers, denial-of-service attacks, etc.
etc.so what you are suggesting is to disable this feature which is already
implemented and in use. Even if this is done ndia wide (no law/regulation
forcing this), what happens outside India? What you suggest is that all
router owners (specially back bone and ISPs physically disable the feature
via software command - for what benefits? This leaves the "Internet" open
to attacks / abuses as there will be no "filter" to block. (That was the
rogue traffic mentioned previously.)

When I send this e-mail, most likely the first packets travel to USA to
determine who "[email protected]" is and where to route packets as final
destination. The one has no idea how my mail goes to arrive at you - some
packets via USA, some via UK and yet another via Germany! That is the
beauty of Internet. Nevermind how once it arrives in India it is routed to
you wherever you are.

Remember all big companies (facebook, google, twitter, youtube, netflix,
etc. etc. have their primary data warehouses outside India (to circumvent
Indian laws) so whenever you use any of these, traffic is generated from
India to wherever these guys data-centers are!

One of the idea - most common one which I mentioned - was to create virtual
circuit switch; ie all packets travel the same way using the same
hops/routers; like in the good-old-telephony.

On 21 May 2015 at 04:53, Arun Mehta <[email protected]> wrote:

> stOn Wed, May 20, 2015 at 2:19 PM, Kiritkumar Lathia <[email protected]>
> wrote:
>
>> *Narendra *wrote: "Dear telco, Please design or create new business
>> model where telco can get the profit with following strict net neutrality.
>> We are ready to pay whatever it cost."
>>
>> COAI argument: OTT (Skype, WhatsApp, ...) have taken away cash cow (voice
>> calls) and so we need to increase data costs by x6 to stay in business !
>> Alternatively, the Telco Regulations should apply to them (charging, call
>> terminations charges, call data storage for x months, police wire-tapping,
>> spectrum costs, universal service obligations, area coverage, ....)
>>
>
> I would like to see their calculations for the 6x -- in any case, setting
> tariffs is the job of TRAI, so I hope they are looking at this "calculation"
>
>
>>
>> Remember: Telco is a regulated industry whereas "Internet" is not
>> regulated at all (there is self regulation but not government regulations).
>>
>> *Arun *had talked about "policing": Here it is not human police but the
>> "technical term" as Arun knows - so replace "policing" by QoS whereby
>> "rouge" packets (corrupted packets, duplicate packets, data heavy packets,
>> ...) are throttled or discarded (!) to provide QoS to all "valid" packets.
>>
>
> Again -- this attempt at improving QoS will only worsen QoS -- it is much,
> much  better to use the money for faster switches, and stuff. Your road
> analogy, Kirit -- reserving a lane for priority traffic certainly slowed
> down the speed for everyone else, but the confusion that it caused at
> crossings led perhaps to even the priority traffic moving slower. Imagine,
> as in the case of optic fiber, that improved technology increses its
> capacity all the time.  For instance, "recent advances in modulation
> technology mean that with QPSK, 4QAM or 16QAM modulation, 1/80th of a dark
> fiber pair can carry a 100 Gbps
> <http://www.cisco.com/en/US/prod/collateral/routers/ps5763/datasheet_c78-478689_0629.html> signal
> in the optical space previously occupied by a single long distance 10 Gbps
> circuit." (
> http://www.metafilter.com/131783/The-data-capacity-of-a-fiber-optic-cable).
> Does it in this case make sense to spend money on technology that in the
> name of QoS slows down traffic, or better switches that increase capacity
> tenfold? Whether you call it policing or QoS, it's a lose-lose.
> Arun Mehta
>

_______________________________________________
India-gii mailing list
India-gii-IAPFreCvJWP2/[email protected]
https://lists.india-gii.org/mailman/listinfo/india-gii