Location might be a tangent, if we should go for just two locations. The
change of unwanted things happening in one location is a too high risk
for an organisation of our importance. The change of unwanted things
happening in two, quit remote, locations happening at the same time,
might be acceptable. But what if we go for three? San Fra (well, we are
there), Amsterdam has great internet connectivity (Amsterdam Internet
Exchange), and Switzerland (or Sweden or Australia) as a third location?
The distribution of the data could be a burden, but if San Fra (for the
time being) is used as main point and information is distributed to the
other two locations with a reasonable delay, we should, in case of a
real disaster, only loose a couple of minutes of saved work if San Fra
would be shut down.
Tests with switching between (the) two locations have been done in the
last year, it was an inconvenience that editing was not possible for
less than an hour, but I found it acceptable. The loss of all our work
is of a really other scale, and definitely not anywhere on the scale of
Fæ schreef op 2019-01-09 12:44:
Location: This is a tangent, one that has been raised before as a
/non-answer/ to the issue of actually getting on with contingency
planning. Realistically I would start by looking at the potential
matches of Sweden, Switzerland, the Netherlands (where servers already
are used for WMF operations), or lastly and for very different
What I find weird, or bizarro, is that the responses so far are vague
dismissals for non-good fantastic reasons, at the level of "let magic
blockchain technology solve it for free", rather than taking on board
that preparing a hot switch for Wikimedia operations in a welcoming
host country, is a highly cost effective disaster contingency plan,
whether due to natural disasters in San Fran / Florida / Amsterdam, or
due to national government using its legal authority to freeze, switch
off or tamper with content due to politically inflated "security" or
"emergency" issues. The risks are real and predictable, and as a
globally recognized charity with plenty of money in the bank, the WMF
should have contingency plans to ensure its continued existence, as
any professional business actuary would advise.
As a past IT auditor, what also made the hairs prick up on the back of
my neck, was David Gerard's sensible question "So ... when did someone
last test putting up a copy of the sites from
the backups" - Could someone give a real answer to that please? If
it's never, then wow, we all have to ask some hard questions of the
WMF Board of exactly how they hold senior management to account.
On Tue, 8 Jan 2019 at 23:05, Nathan wrote:
I'm curious what nation you have in mind for your stable Plan B. Is it
Brexit Britain? France of the Yellow Vests and Front National? Perhaps
Orban's Hungary, Putin's Russia, or Germany with its recent right-wing
Maybe you'd prefer Jair Bolsonaro's Brazil? I suppose in Italy we'd
about Beppe and criminal libel statutes, while BJP would hardly seem
welcoming in India and I can't imagine you'd suggest a home on the
side of the Great Firewall.
Maybe you're hinting at Canada, but otherwise, I'd love to understand
island of liberal stability and legal safeguards you think is safe
vagaries of electoral politics or rigid authoritarianism.
The countries I list above have their own flaws (although in each
believe, many desirable traits as well) as does any other alternative.
Anyone could reasonably argue it's unfair to stigmatize any of them by
glaringly public flaws.
To my mind Steve Walling has it right - the very nature of Wikipedia
maybe the best protection there could be, even against the absurdly
unlikely circumstance of a United States government takeover of
On Tue, Jan 8, 2019 at 12:17 PM Fæ wrote:
> Dear fellow Wikimedians, please sit back for a moment and ponder the
> For those of us not resident in the US, it has been genuinely alarming
> to see highly respected US government archives vanish overnight,
> reference websites go down, and US legislation appear to drift to
> whatever commercial interests have the loudest current political
> voices. Sadly "populism" is happening now, and dominates American
> politics, driving changes of all sorts in response to politically
> inflated and vague rhetoric about "security" and "fakenews". It is not
> inconceivable that a popularist current or future US Government could
> decide to introduce emergency controls over websites like Wikipedia,
> virtually overnight.
> The question of whether the Wikimedia Foundation should have a h