 ONiallYum, CitizenPremium
join:2002-11-18
Portland, OR | i would have picked one comment to reply to... but there are a number of points to raise
- google pays for their access, just like everyone else. their data centers are served by a provider, and google selects the amount of bandwidth they wish to pay for users to access their services.
- the internet is a mesh of carriers, and at certain points there can be failures that affect some remote users, and not others. this can be due to volume, hardware failure, line breaks, or software freak-outs like protocol drops and other wonderful mayhem convincing multiple devices to recognize their co-existence across great physical distances.
- because content providers pay for allocated bandwidth, any fear of youtube or google causing congestion on the network is the result of undercapacity on the part of a carrier, and not the fault of the content provider, except in the instance that the content provider has no more allocated bandwidth available due to utilization. it would be very expensive for a carrier to dedicate backbone bandwidth equal to the summation of the full capacity of all their customers allocated access, so more likely than not, the true bandwidth available will be substantially less. should every user attempt to pull data at the throughput of their allocated bandwidth, the congestion would be immense.
- carriers can build out to provide the summation of all their customers bandwidth, however, you would find that the costs for access would increase to cover that expense.
- the benefits of quality of service, or more specifically, class of service, only is relevant in a situation where network congestion occurs. if there is plenty of bandwidth along the entire source / destination path, then there is no need for packet prioritization. qos and cos are traffic policing features to enable traffic deemed 'critical' to pass through at a higher guaranteed rate than traffic with a lower priority tag.
cheers,
--
i drink with bears for weekend carnivals. |
