[committee] arctic.uniirc.com

[Frames]

I'm not on committee so I'll make no comment.
But I don't see the reason to do this (But on the other hand it is only 
$50).
Like rather than paying $50, we can just spin arctic.uniirc.com up as a 
VM on our hardware -- or not.

IRC servers are easy to link, it is a line in the config of each of the 
servers.
Almost all UCCans on UniIRC do it via UCC's IRC server on mussel.

The only reason this was started was become some individual (Nick Bannon?)
felt it was worth doing. It was not a committee, nor a wheel decided action.

UCC's been hosting our sever for ages longer than this,
and for ages as part of this CASSA's server was linked to it.
And we physically host ComSSA's server (PMS),
so I doubt they would have any problem with us hosting the IRC.

So I'm not seeing any need for AWS hosting.
Sure it isn't much, but it ain't getting us anything.

It is probably work keeping the uniirc domain name paid for though.

But like committee can do what they want.
[*OX]
Wheel member


On 17/08/2016 8:30 PM, Delan Azabani wrote:
> Hi there, committees of CASSA, MITS, and UCC,
>
> I’m writing to you because your clubs are the founding (and as far as
> I can tell, the only) members of UniIRC. Despite our sporadic efforts,
> we haven’t yet acquired any clubs from outside WA, let alone achieved
> world domination (of IRC services in university computing societies).
>
> arctic.uniirc.com is the heart of the UniIRC network — while end users
> are intentionally forbidden from connecting to it directly, every last
> message that we send needs to cross it to reach the rest of us.
>
> After my free year of AWS ran out, ComSSA has paid for arctic for the
> last seventeen months, so I’d like to set up some kind of arrangement
> to share the future costs between our clubs. Before I dive into the
> details — TL;DR: send me $50 each and we’re good for three years.
>
> To give you an idea of our previous costs, last month’s bill for my
> AWS account — which has only ever been used for arctic — was for a
> total of 22.85 AUD, which roughly consists of:
>
> 14.88 USD for 744 hours of t1.micro “on demand”
> + 0.68 USD for 8 GB of storage and half a million I/O requests
> + 0.05 USD in data transfer charges
> × 1.10 GST
> × 1.33 AUD/USD
>
> “But wait!” you might ask. “If we split arctic four ways, $50 each is
> only $200, but arctic would cost over $740, even if every month was
> magically shortened to 28 days and the exchange rate never changed!”
>
> Reserved instances! EC2 instances are “on demand” by default, but if
> we commit to keeping arctic for the next three years, we’ll save 45%
> on the main hourly rate (0.011 USD/h versus 0.02 USD/h).
>
> There are other options available, like a one year commitment, or the
> ability to pay for less than 100% of the reservation upfront, but of
> course, the savings will decrease with those options.
>
> Migrating arctic to a reserved instance means that we’ll need to shut
> it down, take an image of it, destroy the instance, then create a new
> instance with that image, but we have an Elastic IP address, so the
> IP address will *not* need to change.
>
> As a result, we have an opportunity to change arctic from a t1.micro
> (an obsolescent instance type) to a t2.micro (same price, but faster,
> and ~40% more RAM) or a t2.nano (same as a t2.micro, but half the
> price, and ~10% less RAM than a t1.micro).
>
> That said, if we migrate from a t1.micro to a t2.nano, we’ll lose two
> things: about 10% of our RAM, and (like any T2 instance) the ability
> to attach storage to our instance. Neither of those are even remotely
> useful for arctic though — it’s just an IRC server.
>
> I propose that we reserve a t2.nano instance for arctic, for the next
> three years, and pay for it upfront. That’s 159.50 USD after GST, or
> about 212.13 AUD with a 1.33 AUD/USD exchange rate. While it doesn’t
> include charges for storage, I/O, data transfer, and so on, I’m happy
> to eat the costs for those as long as they stay under 2 AUD a month.
>
> If you’re all happy with my proposal, and we all agree on it, then it
> would be ideal if you could send the dough my way by Tuesday evening,
> so that we can start the migration process on Wednesday morning.
>
> A maintenance window of an hour or two should be long enough, and I’ll
> be sure to send out warning notices at appropriate times.
>
> Feel free to send me your questions, complaints, and/or suggestions!
>
> Cheers,
> Delan Azabani
> President
> ComSSA
> _______________________________________________
> List Archives: http://lists.ucc.gu.uwa.edu.au/pipermail/committee

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ucc.gu.uwa.edu.au/pipermail/committee/attachments/20160817/a01ae996/attachment.htm>