RE: Completion time oversight

Karl J. Runge (
Sun, 13 Apr 97 14:41:02 PDT

> >One thing that has been completely ignored in all the estimates
> >of the completion time for this thing is that students go home for
> >the summer in about a month. When that happens the curve will
> >probably stop rising and might even start to fall. Then it should
> >pick up again in September.

We are in a phenomenal rate of growth the past 3-4 days: 20-25% *per day*

If the exponential growth model doesn't bust, from fitting over the last
4 days, we'd be 1/2 done on Wed May 7, and all the way done Sun May 11
with 289080 and 643500 hosts, respectively. Yow!

It will be very interesting to see how "elastic" the current Internet is
to this sort of project. Can it get that high or will we hit "the wall"???

Even fitting over the last 10 days, that has a "modest" 10% growth per
day, shows us 1/2 way thru the keyspace on Sun May 25, and done within
a week after that. (160K and 330K hosts, resp.)

So we just might BE DONE BEFORE Summer vacation. Keep watching ... and keep
spreading the word!

An important issue is will the keyserver be ready for such high rates???

As far as "competing" with the other groups, it may make the difference
to avoid server drag that turns people away to other groups.

I note that DES violation ( ) is at about
186 million keys / sec and

the SolNET group ( ) is at
195 million keys / sec.

We're at 530 million keys /sec, but if you combine the two other groups
they just about equal our production.



PS if anyone want's my perl script that does the least square fitting to
Rocke's desstat.htm page just ask.

Karl J. Runge -- Linux: it's the Real thing --
If you don't reinvent the wheel every now and then (510)-516-7127
how do you know for sure you can invent anything?