Greg Scher gripes about Skype, which is fine, but makes some seriously incorrect assumptions. These were brought to my attention by a reader who pointed out these three key points:
- The article says that after authentication, Skype does not need central infrastructure. This is true only legalistically. Realistically supernodes are part of Skype infrastructure and they could affect the call setup delay while locating the fard-end.
- Secondly it says that because of encryption privacy is protected. This is true only for exchange of voice. I consider my calling patterns – whom I call, who call me – are private information, but by necessity my Supernode knows this information. It is not clear how Skype ensures privacy of this data.
- Thirdly, it says that unlike other VoIP architectures, Skype does not have “SIP gateway” etc and so they are abusing VoIP moniker. The comment about the architecture is flat out wrong and I don’t agree that Skype is not VoIP.
Meanwhile Keith points out that nearly 6000 new downloads an hour for Skype. (See the total downloads at the top of the page. That really doesn’t mean anything, because I would like to know how many actual users. People who use it actively!
Andy Writes: According to Kelly Larabee, Skype’s publicist, she told me on Thursday that out of the 20 some million downloads, that 10 million individuals have downloaded the software. the number that Keith and their site refer to are actual clients logged on. Now, that does not mean users, as for example I have three machines on in the house and each is possibly running skype at the same time….or not….depending if I log off and exit or not.