[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [tor-talk] Speculation: Next 10 years of Tor?

On Thu, Jul 04, 2013 at 02:58:30PM -0400, grarpamp wrote:
> Tor's been around in force for 10+ years.
> In that time it has experienced many things and been used in many ways.
> Where will, or will not, Tor be 10+ years from now, and why?

Tor speeds are interesting. I think they'll become much faster -- fast enough that 
using Tor won't feel slow. Users will flock to it for the privacy, and because it 
will be more than fast enough for general use.

How fast? There's Nielsen's Law, similar to Moore's Law, but for network bandwidth. 
It says that end user connection speeds increase by 50% a year:

Are there metrics on connections speeds for a typical Tor user? I'm not aware of any.
I'm going to speculate and say a typical Tor connection speed today is 100 kpbs. 
Assuming the 50% rule, this becomes 5.8 Mbps in 10 years (100 kbps * 1.5^10). That 
would be more than fast enough for even video. (Video would probably become higher 
bandwidth than today, but hopefully with the kind of resolution options YouTube has.)

Then for hidden services, if they follow the 50% rules as well, a 15 kbps connection 
today would become a 860 kbps connection in 10 years.

Improvements are needed to Tor itself, for this to happen; e.g. there's the call out
for help with hidden services:

Hidden Services need some love

But, the larger improvements, it seems, are going to be with network throughput in 
general. Today a Tor relay that get's 4 Mbps would get 230 Mbps in 10 years, if the 
50% rule applies. (Although, Nielson's Law is for end user connections and not 
server connections. I would expect server connection speed to grow even faster, but 
am not aware of any stats on this...)

tor-talk mailing list