[ih] a single organization "managing the network"

[John Gilmore]

Jack Haverty via Internet-history <internet-history at elists.isoc.org> wrote:
> With all of the corporations now doing "cloud computing", wouldn't
> such a replication be a small blip in their operation that wouldn't
> even be noticed by a Netflix or Youtube or AWS or the like?

Running more servers and spinning drives would be easy for a cloud
vendor to clone.  Keeping up with gigabits/sec of updates: from the big
web crawls and the myriad little ones; and the realtime continuous
digitizing of 50 channels of television; and the scanning in of
thousands of books and magazines daily; and all the public uploads:
would take some good synchronization software, but it's "just engineering".

Providing public access to the collection would be a significant
software undertaking.  If you have the corpus on ten thousand
18-terabyte drives, updated til yesterday, but nobody can access it,
have you really replicated it usefully?

The hard part is building a team of people who manage it all and keep it
running and keep evolving it, as the web and the Internet and the
culture evolve.  And who keep raising the funding, year after year.  And
who do it all independently of the IA, otherwise IA is still a single
point of failure.

	John