Server room clearout - London Hackspace

The old Google Groups will be going away soon, but your browser is incompatible with the new version.
Message from discussion Server room clearout
Received-SPF: pass (google.com: domain of markstew...@gmail.com designates 209.85.214.173 as permitted sender) client-ip=209.85.214.173;
MIME-Version: 1.0
In-Reply-To: <4FDDC490.4090...@phoenixhaven.net>
References: <49fc0a1c-fbce-44d3-8f07-e524bb68b...@b1g2000vbb.googlegroups.com>
	<95fb4f99-eece-4fcd-8da2-5d165752f...@cu1g2000vbb.googlegroups.com>
	<CAFiHz+Z2TYxrKAFgp9mS6qX=epFqzBWqD_B4anStbVr-vjO...@mail.gmail.com>
	<CAEPq4ipYcMJp6ivpYtqOqyKYDEmeSvtgNsAncENe5x3KVdj...@mail.gmail.com>
	<e4a3e7df-6aff-4cc2-8d0f-4706f0355...@n5g2000vbb.googlegroups.com>
	<4FDCC570.3020...@phoenixhaven.net>
	<alpine.DEB.2.00.1206161955360.3079@ybpnyubfg>
	<4FDDC490.4090...@phoenixhaven.net>
Date: Sun, 17 Jun 2012 13:12:25 +0100
Message-ID: <CAPyX2nduWe688u-SdqMOcHzFG1uvwdQMYW61x6CVGQACGKH...@mail.gmail.com>
Subject: Re: [london-hack-space] Re: Server room clearout
From: Mark Steward <markstew...@gmail.com>
To: london-hack-space@googlegroups.com
Cc: Jasper Wallace <jas...@pointless.net>, Simon <skl...@gmail.com>
Content-Type: multipart/alternative; boundary=f46d0447a0b163a5ed04c2a9f623

--f46d0447a0b163a5ed04c2a9f623
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

On Sun, Jun 17, 2012 at 12:50 PM, Nin Lil'izi
<nin-lil-...@phoenixhaven.net>wrote:

>  *>* Any idea what kind of webservices? webmail? forums? different servic=
es *>*
>  will have different loads and so need more or less cpu's or disks.
>
> First stage of the project is some bespokish social networking applicatio=
n. With us being a nation without territory. we were trying to build around=
 a sort of virtual nation theme. Clever marketing maybe. But more of a hook=
 to build the sense of (abeit nationless) identity to the roma nation among=
st the younger generation and entice more of them to become active in fight=
ing to change the poor deal they get in life and to defend their unique cul=
ture.
>
> Why not facebook?
> Because it's far too easy to mine. With most living under regimes of inst=
itutionalised racism with government sponsered gangs of door to door racist=
s. This is very sensitive data. And the whole setup needs to be locked down=
 like a ducks ass after somebody bought a new welder. Also, the other big p=
roblem is language. AFAIK, the first language of most Roma have never been =
localised in any software ever. At least not on the internet or anywhere it=
 counts. So, everything produced would need a unique set of localizations.
> A large number are totally illiterate. So ontop of people wanting to uplo=
ad photo's (User videos some day in the future. But understand the technica=
l challenges here would make this impractical for the forseeable future.) .=
 Any informative content also needs to be available as either audio or vide=
o.
> The next stage involves once people are getting online and interacting in=
troducing demographic elements. So there is a way of guaging what the peopl=
e want on a large scale. And polling for opinions on various issues which a=
ffect directly. So, that the NGO organisations that represent the Roma on t=
he world stage have access to actual representative information about what =
the people want and think. And such data would also provide a spotlight to =
show up the organisations which are using Roma as political bargaining chip=
s rather than being helpful. And strengthen the ones that are doing it righ=
t.
>
> This project is not being undertaken by or the property of any organisati=
on. It is intended to be run as an entirely politically and organisationall=
y neutral entity of it's own. Only opening up polling and collating of data=
 to pro-roma NGO's under a strict set of guidelines (that are still to be d=
rawn up).
> And the project is being worked on by a small group of individuals of Rom=
a heritage. Mainly old skool tribal eldery types, with the help of a few ex=
perts in the other relevant fields that this crosses over into. For some re=
ason this appointments me as queen of all things tech, but ho hum :)
>
> So, don't need multi TB from the very start.... But room for storage redu=
ndancy and the initially tricky to guage initial growth without worrying /t=
oo much/ is where my mind at.
> The only tricky thing that is priority in my mind. Is the high level of u=
ser concurrency. Which is mainly RAM + a godly level of dead reckoning.
> I am considering architecting a few of the backend services as separate V=
M's on the iron from the start.... So that if it comes to expanding and bre=
aking out into extra machines and devices in the future. Then it is an easy=
 and straightforward task.
>
> There is a plan to be applying to places for grants once this gets underw=
ay... But the initial setup and launch is coming out of our own threadbare =
pockets. Yes, I realise how insanely ambitious this is. But where's the fun=
 in life without insurmountable challenges?
>
>
My advice for this is: don't worry.  It's far more work to manage the code
and content than the infrastructure, and that's what will determine whether
you succeed in the first place.  A powerful server is only useful if you're
serving vast amounts of data, or reselling it to people who want their own
(reasonably powerful) servers.

Start off with a small server that will run whatever software you need, and
let the provider worry about redundancy, networking and disaster recovery,
so you can focus on the important things like keeping the setup secure.
 There's no point over provisioning - stuff gets cheaper all the time, and
your money is better spent when you're reasonably sure you'll need it.


Mark

P.S. not sure it changes anything, but there does seem to be a (very small)
Wikipedia at http://rmy.wikipedia.org

--f46d0447a0b163a5ed04c2a9f623
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<div>On Sun, Jun 17, 2012 at 12:50 PM, Nin Lil&#39;izi <span dir=3D"ltr">&l=
t;<a href=3D"mailto:nin-lil-...@phoenixhaven.net" target=3D"_blank">nin-lil=
-...@phoenixhaven.net</a>&gt;</span> wrote:</div><div class=3D"gmail_quote"=
><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1=
px #ccc solid;padding-left:1ex">

 =20
   =20
 =20
  <div bgcolor=3D"#FFFFFF" text=3D"#000000">
    <pre><div class=3D"im"><b>&gt;</b> Any idea what kind of webservices? w=
ebmail? forums? different services=20
<b>&gt;</b></div> will have different loads and so need more or less cpu&#3=
9;s or disks.

First stage of the project is some bespokish social networking application.=
 With us being a nation without territory. we were trying to build around a=
 sort of virtual nation theme. Clever marketing maybe. But more of a hook t=
o build the sense of (abeit nationless) identity to the roma nation amongst=
 the younger generation and entice more of them to become active in fightin=
g to change the poor deal they get in life and to defend their unique cultu=
re.

Why not facebook?
Because it&#39;s far too easy to mine. With most living under regimes of in=
stitutionalised racism with government sponsered gangs of door to door raci=
sts. This is very sensitive data. And the whole setup needs to be locked do=
wn like a ducks ass after somebody bought a new welder. Also, the other big=
 problem is language. AFAIK, the first language of most Roma have never bee=
n localised in any software ever. At least not on the internet or anywhere =
it counts. So, everything produced would need a unique set of localizations=
.
A large number are totally illiterate. So ontop of people wanting to upload=
 photo&#39;s (User videos some day in the future. But understand the techni=
cal challenges here would make this impractical for the forseeable future.)=
 . Any informative content also needs to be available as either audio or vi=
deo.
The next stage involves once people are getting online and interacting intr=
oducing demographic elements. So there is a way of guaging what the people =
want on a large scale. And polling for opinions on various issues which aff=
ect directly. So, that the NGO organisations that represent the Roma on the=
 world stage have access to actual representative information about what th=
e people want and think. And such data would also provide a spotlight to sh=
ow up the organisations which are using Roma as political bargaining chips =
rather than being helpful. And strengthen the ones that are doing it right.

This project is not being undertaken by or the property of any organisation=
. It is intended to be run as an entirely politically and organisationally =
neutral entity of it&#39;s own. Only opening up polling and collating of da=
ta to pro-roma NGO&#39;s under a strict set of guidelines (that are still t=
o be drawn up).
And the project is being worked on by a small group of individuals of Roma =
heritage. Mainly old skool tribal eldery types, with the help of a few expe=
rts in the other relevant fields that this crosses over into. For some reas=
on this appointments me as queen of all things tech, but ho hum :)

So, don&#39;t need multi TB from the very start.... But room for storage re=
dundancy and the initially tricky to guage initial growth without worrying =
/too much/ is where my mind at.
The only tricky thing that is priority in my mind. Is the high level of use=
r concurrency. Which is mainly RAM + a godly level of dead reckoning.
I am considering architecting a few of the backend services as separate VM&=
#39;s on the iron from the start.... So that if it comes to expanding and b=
reaking out into extra machines and devices in the future. Then it is an ea=
sy and straightforward task.

There is a plan to be applying to places for grants once this gets underway=
... But the initial setup and launch is coming out of our own threadbare po=
ckets. Yes, I realise how insanely ambitious this is. But where&#39;s the f=
un in life without insurmountable challenges?
</pre>
    <div><div class=3D"im"></div></div></div></blockquote></div><br><div>My=
 advice for this is: don&#39;t worry. =A0It&#39;s far more work to manage t=
he code and content than the infrastructure, and that&#39;s what will deter=
mine whether you succeed in the first place. =A0A powerful server is only u=
seful if you&#39;re serving vast amounts of data, or reselling it to people=
 who want their own (reasonably powerful) servers.</div>
<div><br></div><div>Start off with a small server that will run whatever so=
ftware you need, and let the provider worry about redundancy, networking an=
d disaster recovery, so you can focus on the important things like keeping =
the setup secure. =A0There&#39;s no point over provisioning - stuff gets ch=
eaper all the time, and your money is better spent when you&#39;re reasonab=
ly sure you&#39;ll need it.</div>
<div><br></div><div><br></div><div>Mark</div><div><br></div><div>P.S. not s=
ure it changes anything, but there does seem to be a (very small) Wikipedia=
 at=A0<a href=3D"http://rmy.wikipedia.org/">http://rmy.wikipedia.org</a></d=
iv>

--f46d0447a0b163a5ed04c2a9f623--
Create a group - Google Groups - Google Home - Terms of Service - Privacy Policy
Account Options
