What do you mean by a local resolver? If you're talking about a local
handle resolver adhering to the handle spec... well, then it depends on
the handle spec I guess, which I don't know. But since all the handle
documetnation keeps saying "like DNS", then I'd imagine it has similar
(or better) redundancy built into it as DNS does. But I don't know.
Poking around on handle.net, it looks like the handle infrastructure
supports this,but you would have had to actually configure 'backup'
handle resolvers -- similar to DNS in that if the DNS for your domain
goes down, and you _haven't_ gotten someone else at another location to
be a 'backup' resolver for you, and specified them as a nameserver in
your DNS record... then you're out of luck. But the protocol supports
that, and if you have done it (as most everyone does with DNS), you're
I have no idea if 'most everyone' does it with handle or not, but handle
supports it. Note that if dx.doi.org goes down, you obviously won't be
able to resolve at dx.doi.org -- but IF it works as I think (I'm still
confused), AND dx.doi.org has distributed their handles to a backup
resolver, then you'd still be able to resolve via hdl.handle.net, or via
your own local handle resolver (which will in turn find the backup
Ben O'Steen wrote:
> What happens if the main doi resolver goes down? I'd be interested to see
> how well a local resolver works when blocked from this upstream server. Are
> there any other upstream servers?
> On Nov 23, 2009 10:10 PM, "Tom Keays" <[log in to unmask]> wrote:
> Interesting stuff. I never really thought about it before that DOIs
> can be served up by the Handle server. E.G.,
> http://dx.doi.org/10.1074/jbc.M004545200 <=>
> But, even more surprising to me was realizing that Handles can be
> resolved by the DOI server. Or presumably any DOI server.
> http://hdl.handle.net/2027.42/46087 <=> http://dx.doi.org/2027.42/46087
> I suppose I should have understood this point since the Handle service
> does sort of obliquely say this.
> Anyway, good to have it made explicit.
> On Mon, Nov 23, 2009 at 4:03 PM, Jonathan Rochkind <[log in to unmask]> wrote:
>> The actual "handle" ...