Lazlow
join:2006-08-07
Saint Louis, MO | reply to openbox9
Re: Not Shocking, but "This observation supports the assertion that the heaviest users over a month only have a marginal impact on peak-time traffic levels."
Since ISP costs are based on peak time usage(both hardware and transit) this shows that the heaviest users are not causing a significant cost increase. |
|
openbox9
join:2004-01-26
Alexandria, VA
kudos:2 | Maybe I'm misreading the the rest of that paragraph, but how does Sandvine draw that assertion? The same data also reveals that as the observation window decreases, the total percentage of bytes attributable to heavier network users increases. For instance, the consumption kings for any particular day account for 30% of total bytes, and represent a distinct (but over-lapping) set of subscribers when compared against the consumption kings for a month or a week. Similarly, past research conducted by Sandvine has demonstrated that in any peak time hour, the top users during that hour are responsible for an even higher percentage of total bandwidth, but again are largely distinct from the consumption kings of the preceding month. Distinct but over-lapping set of subscribers? |
|
| Anyone who trusts "statistics" from a company who has a vested interest in those "statistics" is a fool. |
|
|
|
Lazlow
join:2006-08-07
Saint Louis, MO | reply to openbox9
Without the data I suspect that it would be hard to tell how they do it for sure. My suspicion would be that the highest(GB/month) users are probably on lower speed teirs (ie 5,10Mbps) and thus even when downloading at full speed have minimal impact on aggregate peak Mbps. |
|
openbox9
join:2004-01-26
Alexandria, VA
kudos:2 | Agreed. It's difficult to draw serious conclusions without more information...especially based on that paragraph that seems contradictory. |
|
