Experimental Physics and
Industrial Control System

Torsten Bögershausen via Core-talk <core-talk at aps.anl.gov> · Wed, 17 Mar 2021 13:30:38 +0000

Hm, why is this so complicated ?
Because it is.

For my example, I didn't care what a switch really is.
A switch is where you plug in a bunch of Ethernet cables, and may be optical fibers theses days.
And yes, there are switches and switches, but that is not the point at the moment.

The paper says "IPv6 only", right ?
In order to get an idea what needs to be done, I imagined that the switches don't transport
IPV4 any more.
Those switches do probably not exist in reality. You need to imagine them.
There are not datasheets, they do not exist.

Imagine them.

What needs to be done ?
Each and every "packet" needs to be transported through IPv6,
probably UDP and/or TCP.

What needs to be done in terms of SW development ?

My apologies  if I waste your time with imaginations.

On 2021-03-17, 14:16, "Timo Korhonen" <Timo.Korhonen at ess.eu> wrote:

    What exactly is meant with a "switch"? For instance, Cisco talks about "Layer 2" and "Layer 3" switches:

    https://documentation.meraki.com/MS/Layer_3_Switching/Layer_3_vs_Layer_2_Switching

    Layer 2 switch is what Mark is referring to, and should not even know the difference between IPv4 and IPv6. Layer 3 switch is a "router" which will know the difference.

    Timo 

    On 17/03/21 13:59, "Core-talk on behalf of Mark Rivers via Core-talk" <core-talk-bounces at aps.anl.gov on behalf of core-talk at aps.anl.gov> wrote:

        > Imagine, that one day all switches at APS start to refuse to transport IPv4, they only transport IPv6.

        I still don't understand.  Switches do not know what protocol they are transporting, they just route low-level Ethernet packets based on the destination MAC address.  They can transport protocols like IEEE 802.2 SNAP, which some of our devices use.  That is not an IP protocol at all.  I don't believe switches will ever refuse to transport IPv4.

        Mark

        ________________________________
        From: Torsten Bögershausen <torsten.bogershausen at ess.eu>
        Sent: Wednesday, March 17, 2021 7:50 AM
        To: Mark Rivers; 'Michael Davidsaver'; Johnson, Andrew N.; Ben Franksen
        Cc: EPICS core-talk
        Subject: Re: IPv6

        OK, OK.

        I should have written:

        Imagine,
        that one day all switches at APS start to refuse to transport IPv4,
        they only transport IPv6.

        Which SW adaptions would be needed to get a working facility ?
        With EPICS and all the EPICS-related tools working as today.
        (And all non-EPICS stuff).

        That question to ask is easy, to give an answer is not easy.

        /Torsten

        On 3/17/21 1:19 PM, Mark Rivers wrote:
        >> Think that all switches, routers, firewalls refuse to transport IPv4 one day.
        >
        >
        > I believe that switches operate below the level of the network protocol, they just operate at the level of Ethernet packets. I don't think they can refuse to transport IPv4.  They are currently transporting some traffic that is not even IP at all.
        >
        >
        > Today all of the traffic between all of my IOCs and devices only goes through switches, no routers or firewalls.  So that should continue to operate fine as long as the OS continues to support IPv4 on the NIC?
        >
        > I don't believe it is necessary to have a local cable.
        >
        > Mark
        >
        > ________________________________
        > From: Torsten Bögershausen <Torsten.Bogershausen at ess.eu>
        > Sent: Wednesday, March 17, 2021 2:45 AM
        > To: Mark Rivers; 'Michael Davidsaver'; Johnson, Andrew N.; Ben Franksen
        > Cc: EPICS core-talk
        > Subject: Re: IPv6
        >
        > My understanding is that the core network must be running IPv6.
        >
        > That is all IOC-IOC communication, channel access, pv access.
        > All clients talking to an IOC:
        > CSS, caget, pvget, archiver, alarm handler, EDM, MEDM and so on.
        >
        > Think that all routes and switches between your office and the beamline
        > onöy transport IPv6.
        >
        > But, back to asyn:
        > To my understanding, once you have an IOC running this way,
        > I think this IOC can "speak" RS232, RS485, USB, Modbus, or even TCP v4
        > to a local device.
        > Think that all switches, routers, firewalls refuse to transport IPv4 one day.
        > You can still have a local cable between the IOC and the device.
        >
        > /Torsten
        >
        > On 2021-03-16, 20:37, "Core-talk on behalf of Mark Rivers via Core-talk" <core-talk-bounces at aps.anl.gov on behalf of core-talk at aps.anl.gov> wrote:
        >
        >      How will we handle the fact that tens of thousands of devices use IP4 and will never have their firmware updated?  Will we still be able to talk to them?  Will their use require an exception?
        >
        >      -----Original Message-----
        >      From: Core-talk <core-talk-bounces at aps.anl.gov> On Behalf Of Michael Davidsaver via Core-talk
        >      Sent: Tuesday, March 16, 2021 2:04 PM
        >      To: Johnson, Andrew N. <anj at anl.gov>; Ben Franksen <benjamin.franksen at helmholtz-berlin.de>
        >      Cc: core-talk at aps.anl.gov
        >      Subject: Re: IPv6
        >
        >      Personally, I see IPv6 as one of several "modernization" requirements which seem likely drop at some point within say 5 years.
        >
        >      The biggest one being something along the lines of "all network services must employ strong authentication and encryption".  I expect that the continuing drumbeat of headlines about PLC security problems will at some point spill over into the EPICS world in a way which makes arguing for exemptions untenable.
        >
        >      Our community can either be proactive, or wait to be surprised.
        >
        >
        >      On 3/16/21 11:33 AM, Johnson, Andrew N. via Core-talk wrote:
        >      > On Mar 16, 2021, at 4:32 AM, Ben Franksen <benjamin.franksen at helmholtz-berlin.de> wrote:
        >      >>
        >      >> Am 16.03.21 um 09:44 schrieb Zimoch Dirk (PSI) via Core-talk:
        >      >>> On Fri, 2021-03-12 at 03:34 +0000, Johnson, Andrew N. via Core-talk wrote:
        >      >>>> https://www.whitehouse.gov/wp-content/uploads/2020/11/M-21-07.pdf
        >      >>>>
        >      >>>> - Andrew
        >      >>>>
        >      >>>
        >      >>> TL;DR
        >      >>> Do we have a problem?
        >      >>
        >      >> I may be wrong, but AFAIU only facilities in the US have a problem.
        >      >> They need to convince their over-bosses that they get an exception.
        >      >
        >      > Sorry, but if we don’t consider how we can add support for IPv6 soon EPICS will probably no longer be eligible for use by the kinds of large experimental facilities that have funded its development to date, and it will die. I’m not saying it’s urgent, but we should start to plan for it.
        >      >
        >      > That OMB memo was signed by the previous US administration, but I’m pretty sure it wasn’t developed by their political appointees, and IMHO hoping that the new administration will rescind it would be a mistake. Existing DOE facilities will almost certainly be getting exemptions of some kind, but EPICS doesn’t have a monopoly in this field and if we don’t support it many future Government-funded projects will have to find an alternative since IPv6-only networking will most likely become a non-negotiable requirement at some point. This isn’t likely to be unique to the US either.
        >      >
        >      > IPv6 use has been growing and many cellphone networks now depend on it (that DJB article which Ben found has a last-modified date of August 2003). There’s a page with links to several adoption statistics websites at
        >      >     https://www.internetsociety.org/deploy360/ipv6/statistics/
        >      >
        >      > I’m hoping that we’ll be able to get some DOE funding to actually do the porting work. Given the number of DOE facilities that use EPICS it seemed reasonable to suggest that, which I have done to the team that is planning the DOE’s response to that memo.
        >      >
        >      > - Andrew
        >      >
        >
        >
        >

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System