[AusNOG] urlscan.io

[Mark Smith]

Read my second paragraph again.



On 16 June 2017 at 17:10, Damian Guppy <the.damo at gmail.com> wrote:
> Then you should know what happens to a reverse proxy when you have a cache
> miss, the request doesn't fail, it just gets transparently proxied off to
> the upstream server that does have the content.
>
> --Damian
>
> On Fri, Jun 16, 2017 at 3:07 PM, Mark Smith <markzzzsmith at gmail.com> wrote:
>>
>> On 16 June 2017 at 16:35, Damian Guppy <the.damo at gmail.com> wrote:
>> > Akamai is a caching network. DNS does not provide the sort of
>> > intelligence
>> > necessary to direct requests to the most appropriate server, so you will
>> > always just hit the server closest to you. If that server happens to
>> > have
>> > the content already cached then it will serve it up itself. If it
>> > doesn't
>> > have some or all of the content required (cache miss) the server will
>> > act as
>> > a proxy and fetch the content from the closest upstream server on the
>> > akamai
>> > network that does have the content, and then hold onto it for an amount
>> > of
>> > time as defined by their internal algorithms in case anyone else needs
>> > that
>> > content.
>> >
>> > *this is a simplified explanation of what is a complex system*
>>
>> You might want to look up the the expression "Teaching grandmother to
>> suck eggs".
>>
>> I've worked on the assumption that CDNs, given the amount of storage
>> that CDN nodes have (I've seen a few in person), and that ISPs have
>> had their own translucent web proxies in the past, have content
>> proactively replicated to them so that there is no latency, even for
>> the first request.
>>
>> >
>> > Thanks
>> > --Damian
>> >
>> > On Fri, Jun 16, 2017 at 2:29 PM, Tim Raphael <raphael.timothy at gmail.com>
>> > wrote:
>> >>
>> >> Mark,
>> >>
>> >> You’ll find that Akamai’s algorithms will retrieve the content from the
>> >> origin and keep it at varying stages of “warm” in their caches based on
>> >> demand.
>> >>
>> >> I’d be pretty unimpressed if I was a US / EU journo trying to get
>> >> Australian news from a webpage 500+ms RT away.
>> >>
>> >> - Tim
>> >>
>> >>
>> >> > On 16 Jun 2017, at 4:25 pm, Mark Smith <markzzzsmith at gmail.com>
>> >> > wrote:
>> >> >
>> >> > On 16 June 2017 at 16:10, Scott Howard <scott at doc.net.au> wrote:
>> >> >> On Thu, Jun 15, 2017 at 10:57 PM, Mark Smith
>> >> >> <markzzzsmith at gmail.com>
>> >> >> wrote:
>> >> >>>
>> >> >>> I think an interesting example is www.theage.com.au. You would
>> >> >>> expect
>> >> >>> the main site to be hosted somewhere inside Australia, yet it is
>> >> >>> being
>> >> >>> hosted by Akamai somewhere in Europe.
>> >> >>
>> >> >>
>> >> >> Want to think about that comment a little more?
>> >> >>
>> >> >
>> >> > Not really, didn't think much about it before.
>> >> >
>> >> > Perhaps it it is surprising that Akamai are hosting copies of content
>> >> > a long way away from where it is going to be popularly read. There
>> >> > can't be that many readers of The Age in Europe.
>> >> >
>> >> > I don't know anything about Akamai's service optons, and whether
>> >> > customers can choose where their content is held or provide an
>> >> > indication of where the content is most likely consumed.
>> >> >
>> >> > If not, it might indicate Akamai's replication strategy could be copy
>> >> > everything everywhere or perhaps at least one copy in each continent.
>> >> > Cheap enough to do in terms of storage and network bandwidth, just a
>> >> > bit of a surprise it isn't more optimal.
>> >> >
>> >> >> Where do you think urlscan.io is hosted?  How does Akamai work?
>> >> >>
>> >> >>  Scott
>> >> > _______________________________________________
>> >> > AusNOG mailing list
>> >> > AusNOG at lists.ausnog.net
>> >> > http://lists.ausnog.net/mailman/listinfo/ausnog
>> >>
>> >> _______________________________________________
>> >> AusNOG mailing list
>> >> AusNOG at lists.ausnog.net
>> >> http://lists.ausnog.net/mailman/listinfo/ausnog
>> >
>> >
>
>