Marc A. Pelletier wrote:
... as far as I know, high-end networking hardware is not available with Libre OSes
Are the FreeBSD-based pfSense C2758 series in the Foundation's throughput tier?
https://www.pfsense.org/products/product-family.html#c2758
https://portal.pfsense.org/docs/manuals/c2758/system-specification.html
https://en.wikipedia.org/wiki/PfSense
What are the current Foundation throughput bandwidth requirements?
Le 22/03/2016 07:04, James Salsman a écrit :
... as far as I know, high-end networking hardware is not available with Libre OSes
Are the FreeBSD-based pfSense C2758 series in the Foundation's throughput tier?
https://www.pfsense.org/products/product-family.html#c2758
https://portal.pfsense.org/docs/manuals/c2758/system-specification.html
https://en.wikipedia.org/wiki/PfSense
What are the current Foundation throughput bandwidth requirements?
I would not qualify the PfSense product as high-end. It is basically a PC with packet management handled at the software layer. Wikimedia does not have FreeBSD systems AFAIK and the operations people dealing with networking would need a training for PfSense.
From wikitech, ones can get a list of hardware routers and switches
being used: https://wikitech.wikimedia.org/wiki/Common_Datacenter_Specifications#EQIAD
They are Juniper, a leader in networking equipments (another is Cisco). Its operating system administration commands are well known by network engineers around the world. That makes it easier to enroll new network people.
According to the wikitech page, the routers are MX80 and MX480 and switches EX4200 / EX4550. They come with integrated circuits to deal with packets, ie it is a hardware chip dealing with packets and network flow. That makes them order of magnitude faster. The Juniper operating system is BSD based and comes with a wide range of features that are imho unmatched in the Libre world.
I understand the idea behind pushing for 100% FOSS, but that should not be a goal of the foundation. As long as the projects can be cloned and rebuild based on FOSS, I think it is good enough.
Surely, I don't see the Foundation asking for buildings plans to be under a creative common or forbid use of Mac OS, Windows or iPhone ? That is really a different goal than sharing knowledge.
Hoi, Given that a Google and Facebook are working on networking equipment and are making both hardware and associated freely available, it makes more sense to concentrate on this high end solution. A solution that fits with the need of WMF. Thanks, GerardM
PS I am not so interested that I know what license it is. I know it is free.
On 22 March 2016 at 11:56, Antoine Musso hashar+wmf@free.fr wrote:
Le 22/03/2016 07:04, James Salsman a écrit :
... as far as I know, high-end networking hardware is not available with Libre OSes
Are the FreeBSD-based pfSense C2758 series in the Foundation's
throughput tier?
https://www.pfsense.org/products/product-family.html#c2758
https://portal.pfsense.org/docs/manuals/c2758/system-specification.html
https://en.wikipedia.org/wiki/PfSense
What are the current Foundation throughput bandwidth requirements?
I would not qualify the PfSense product as high-end. It is basically a PC with packet management handled at the software layer. Wikimedia does not have FreeBSD systems AFAIK and the operations people dealing with networking would need a training for PfSense.
From wikitech, ones can get a list of hardware routers and switches being used:
https://wikitech.wikimedia.org/wiki/Common_Datacenter_Specifications#EQIAD
They are Juniper, a leader in networking equipments (another is Cisco). Its operating system administration commands are well known by network engineers around the world. That makes it easier to enroll new network people.
According to the wikitech page, the routers are MX80 and MX480 and switches EX4200 / EX4550. They come with integrated circuits to deal with packets, ie it is a hardware chip dealing with packets and network flow. That makes them order of magnitude faster. The Juniper operating system is BSD based and comes with a wide range of features that are imho unmatched in the Libre world.
I understand the idea behind pushing for 100% FOSS, but that should not be a goal of the foundation. As long as the projects can be cloned and rebuild based on FOSS, I think it is good enough.
Surely, I don't see the Foundation asking for buildings plans to be under a creative common or forbid use of Mac OS, Windows or iPhone ? That is really a different goal than sharing knowledge.
-- Antoine "hashar" Musso
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines New messages to: Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe
On 2016-03-22 2:04 AM, James Salsman wrote:
Are the FreeBSD-based pfSense C2758 series in the Foundation's throughput tier? [...]
That looks like decent mid-range gear, but definitely not the the hardware-supported levels needed to support operations.
What are the current Foundation throughput bandwidth requirements?
Big. :-) I'm not the best one to ask for those details as I've had very little involvement in the networking side of ops; Faidon or Mark would be your best bet there.
But, to be honest, I think this is besides the point: I'm not arguing that specific piece of gear X or Y needs or should not be replaced with a possible FLOSS-only alternative; but that *attempting* to do so is a difficult, expensive, and manpower-hungry endeavor whether you succeed or not.
There are things where that investment is worthwhile - or even necessary. There are other things where doing so is at best a waste of donors' money (especially for one-offs or accessory parts of what the Foundation does that impact how the work is done rather than the projects).
A good example might be our videoconferencing software. The Foundation uses Google Hangouts a lot. Nowadays, for bigger meetings, Bluejeans has been added to the list. At (very) regular interval, someone in engineering does another round of testing of FLOSS videoconferencing alternatives, because it irks many of us that we rely on proprietary solutions - and every time to date the result is that none will work to cover our use cases properly.
In the end, there are three only choices: (a) pick an inexpensive proprietary solution that does the job, (b) make our own (or participate in making it), or (c) do without. When doing without would prevent the staff from doing the job, that doesn't leave very many options.
-- Coren / Marc
wikimedia-l@lists.wikimedia.org