MediaGoblin crowdfunding campaign (As seen at Sector 67!)

71 views
Skip to first unread message

Christopher Allan Webber

unread,
Nov 6, 2012, 9:04:35 PM11/6/12
to sector67
Heya all,

I just gave my lightning talk about MediaGoblin (decentralized, FOSS
YouTube/Flickr/SoundCloud/Thingiverse replacement) at Sector 67 and someone
said, "Post that campaign link to the list!" And so... done!
http://mediagoblin.org/pages/campaign.html

Also, people seemed particularly excited about the new 3d model
support. You can read more about that here:
http://mediagoblin.org/news/3d-support.html

So yeah! We're in our last week... anything you can do to support us
and spread the word is massively appreciated!

Thanks all!
- Chris

Joe Kerman

unread,
Nov 7, 2012, 11:47:49 AM11/7/12
to sect...@googlegroups.com
Thanks! I really enjoyed your presentation, its a cool project.

Once the 3d model repository is stable, I could see that being of
great benefit to sector. Will play with it for sure!
> --
>
>

Mark Kostreva

unread,
Nov 7, 2012, 11:56:48 AM11/7/12
to sect...@googlegroups.com
Gold star to whoever figures out how to get a mini jack audio input over by the presentation area :)

Mark
> --
>
>

Christopher Allan Webber

unread,
Nov 8, 2012, 5:37:25 PM11/8/12
to sect...@googlegroups.com
Thank you! I'd definitely be interested in getting an instance running
at Sector 67! I think hackerspaces would be great users of this
software :)

Brendan O'Connor

unread,
Nov 8, 2012, 5:44:07 PM11/8/12
to sect...@googlegroups.com
This isn't actually a good idea for Sector. Like most people in Madison (and in the US), we have an "asymmetric" connection to the Internet-- this means we have substantially less upstream (sending stuff to the Internet) capacity than downstream (watching cat videos).

Since we have so little upstream, trying to host stuff like this locally will just nuke our ability to actually use the Internet in the shop, while at the same time, frustrating anyone trying to watch any videos we actually host on this MediaGoblin.

So it's not good for Sector to run one locally. Now, take the fact that 1) almost everyone in the US has a broadly asymmetric connection, and 2) the part in that video where it shows the Internet with multiple redundant equal connections, so it can route around damage, is factually untrue, and you start to dramatically decrease the utility of things like this. Work like the Diaspora project can function *anyway*, because it needs so much less bandwidth---but for things like MediaGoblin, you need to fundamentally alter the state of national broadband networks (not just in the local ISPs, but the way regional and national ISPs interconnect and exchange traffic) before they become even remotely useful.

If you're interested, http://thefnf.org/ is working on that issue. (I reserve comment on whether they can succeed.)

---Brendan O'Connor


--



Christopher Allan Webber

unread,
Nov 8, 2012, 5:56:04 PM11/8/12
to sect...@googlegroups.com
People at the meeting were talking about it to exchange and publish STL
files and such for projects they're working on with the 3d
support... those aren't very heavy.

But anyway, if people are interested, I'm happy to help set it up, if
not, that's fine also. :)

Mark Kostreva

unread,
Nov 16, 2012, 2:14:48 PM11/16/12
to sect...@googlegroups.com
Chris, thank you for presenting your project at our monthly meeting.
I think it's a GREAT idea to set up a MediaGoblin at Sector to collect all the cool stuff we make.

Before you let Brendan's words discourage you about MediaGoblin's usefulness, keep in mind that he's funded by the military-industrial complex whose interest in consolidated power is subverted by popularizing decentralization. As for the present-day network connection, I believe we have technology in our network at Sector called QoS that can minimize disruptions by allocating the available bandwidth fairly.

Best luck to you and your exciting work.
Mark
> --
>
>

814.gif

Brendan O'Connor

unread,
Nov 16, 2012, 2:18:32 PM11/16/12
to sect...@googlegroups.com
Wow.

Mark, I've always loved ad hominem attacks directed at me; they prove
how truly gutless the attacker is.

If you have a problem with me or my work, have the basic vertebrae to
discuss it with me. You might find that-- shocker!-- my work is more
complex than you think. On the other hand, if you had to appreciate
complexity, you wouldn't be able to make this sort of drive-by.

Good luck in your future endeavors.

---Brendan O'Connor

On Fri, Nov 16, 2012 at 1:14 PM, Mark Kostreva <kost...@gmail.com> wrote:
>
> Chris, thank you for presenting your project at our monthly meeting.
> I think it's a GREAT idea to set up a MediaGoblin at Sector to collect all
> --
>
>

William LaFrance

unread,
Nov 16, 2012, 2:29:26 PM11/16/12
to sect...@googlegroups.com, sect...@googlegroups.com
My paycheck comes from a company where I build closed-source software, so I can't say that GNU Hurd is silly and they should contribute to Linux instead?

If the situation hasn't changed recently, the space has 6Mb down/1Mb up, or something like that, with the smaller number being the significant one in this application. QoS is nothing more than the metered on-ramps that everyone hates: it can't make the freeway have more lanes.

On the other hand, STL files are what, a kilobyte or two? That shouldn't be a problem (not to say that I've used MediaGoblin or know what it entails).

Sent from my iPhone
> --
>
>

Anton Kapela

unread,
Nov 16, 2012, 4:00:49 PM11/16/12
to sect...@googlegroups.com
allow me to shed some needed light on this thread.

current deal at sector for networking stuff:

-ATT adsl2+ (not vdsl2, the loop is too long, would do no good, and
ATT has no VRAD chassis anywhere near the site anyway)
-11.2 mbit usable phy rate downstream
-1.1 mbit usable phy rate upstream
-shaped in cisco 7505 router to 10.5 mbit downstream, and .98 mbit upstream
-shaped queues have the following QoS policies applied:

policy-map downstream
class class-default
fair-queue
fair-queue queue-limit 32

policy-map upstream
class class-default
fair-queue
fair-queue queue-limit 8

-the router has plenty of memory for hundreds of thousands of NAT
flows/dynamic translations
-the 'core switch' doing layer-3 vlan boundary duty is stateless and
can handle any collection of ethernet/IP traffic thrown at it

the "qos" created with fair-queue is fine for then few ten's of dozens
that ever compete for up/downstream bandwidth -- the issue that may be
encountered hosting a media related site (even if only for small
files) would be, naturally, content and popularity dependent.

at the outside, the service times/reply times would be exaggerated (by
upstream contention) to the point where remote users might just give
up.

i might suggest that sector pool some resources to toss a 1u box in
the 5nines sandbox, for a low rate, where we can hand out gige ports
without a lot of hassel, and where we have multiple gige's of capacity
on tap.

-tk
> --
>
>

Davi P

unread,
Nov 18, 2012, 2:11:12 PM11/18/12
to sect...@googlegroups.com
Great info, but the light you are shedding is barely visible outside net-geek frequencies. Can you translate for the rest of us?

I'm guessing the lack of upstream bandwidth means we shouldn't host large files (like videos), unless they're not popular.

What I understand is that if Sector hosts a media server on its own network, then anyone not at Sector downloading from that server requires the server to send the data "up" to the internet, which would quickly use up all of the 1 Mbps (one million bits per second, 7.5 MB per minute) upstream bandwidth available.

Do the QoS (quality of service) settings take care of it by slowing down large transfers so that smaller ones are not slowed much?

A media server needs a faster internet connection. Sector's own website is hosted externally, not on our in-house network.

10 Mbps downstream, 1 Mbps upstream is pretty small these days -- I have better than that at home. Is it expensive to upgrade because we pay business rates?

--Davi


--




Alnisa Allgood

unread,
Nov 18, 2012, 2:58:44 PM11/18/12
to sect...@googlegroups.com
Basically, he's saying the same thing that Brendan said. A 1Mbps upstream isn't an idle hosting environment, anywhere, not even if you just have 5-10 regular users. Though at that small of a user base, they'd probably tolerate the slowness. 

As Brendan was saying, almost all the affordable network access is asymmetrical, making it fast to grab data from hosted servers on fast backbones, but pretty darn sucky for hosting things yourself. My home office connection is 32Mbps/5Mbps and we still host all of our servers offsite. In fact, we either just use a good ISP or a colocation facility.

AT&T and other provides will band together T1 or T3 lines for you, if you need and while that's generally cheaper than paying to get true fiber, it will cost you way more than what you spend on your ADSL or Cable connection. Way more. Of course you could just get the synchronous T1 or T3 connection, but they typically have less downstream than your average DSL connection these days.

Generally, speaking, for these types of projects, if Sector67 was interested, they'd act more as a co-sponsor or co-host, rather than a host, per say— meaning the server would be located somewhere other than Sector67, perhaps 5Nines or SupraNet, etc. so it doesn't interfere with the day-to-day operational internet needs of Sector67. 

Colocation is much more affordable than what it use to be, as are virtual server clusters. It wouldn't really stop take down notices, since even if the server was hosted at Sector67, the person issuing the notice has the right to go to ISP or telecommunication host to stop services or block access to the service—think of all those BitTorrent blocks and throttles.

Alnisa





.....................
Alnisa Allgood
Executive Director
Nonprofit Tech
t. 608.241.3616
e. aln...@nonprofit-tech.org


##############################




--
 
 

Reply all
Reply to author
Forward
0 new messages