The graphing is public, there's currently no alerting. From the IX, perspective the only leverage is removal from the fabric or total depeering. Not a very good options. Members can always filter or preference their routes using the MICE communities or policies on their own gear. When we start tail dropping (or getting full) I reach out to the peer directly and make them aware of the situation and offer up MICE upgrade options. Sometimes multiple times. MICE is not advocating members congest their ports, it defeats the purpose of the IX. MICE is not disconnecting member connected congested ports. On Wed, Aug 21, 2019 at 11:28 AM Ben Wiechman <ben.wiechman@arvig.com> wrote:
I would dispute the "only their network" statement. We've troubleshot various issues in the past with latency and loss, especially for internet based services used by businesses when that traffic traverses MICE. While the peering policy of a service provider is their decision, it is potentially appropriate to provide proactive notice in some fashion that would alert other operators that they may want to adjust their peering strategy to minimize customer complaints.
Is some kind of alarming or reporting viable and useful? From my perspective it would be useful.
Ben Wiechman
Director of IP Strategy and Engineering
320.247.3224 | ben.wiechman@arvig.com
Arvig | 224 East Main Street | Melrose, MN 56352 | arvig.com
On Wed, Aug 21, 2019 at 7:39 AM Jay Hanke <jayhanke@southfront.io> wrote:
The board talked about this back in the day. The thought process was that remotes affect multiple members so the congestion policy should be enforced.
For a participant it's only their network (and their customers). Also, a disconnect might make the overall internet worse as their transit may fill.
On Tue, Aug 20, 2019, 11:57 PM AnthonyAnderberg@nuvera.net <AnthonyAnderberg@nuvera.net> wrote:
I know in the past several of us have reached out privately to folks at DCN about their port saturation, which seems to come and go.
We do not have a policy about member-port saturation, and my recollection is similar to Richard's - we didn't want to be bossy about people's peering.
Cheers, anthony
On 8/20/19, 9:05 PM, "Richard Laager" <rlaager@WIKTEL.COM> wrote:
On 8/20/19 8:48 PM, Darin Steffl wrote: > Is there any policy in place for peers that let their ports saturate at > 100% for an extended period of time? DCN looks like they could use an > upgrade to their 10G port and not sure if anyone proactively reaches out > to members when saturation occurs.
I've forwarded your message to DCN.
For remote switch ports, we have a policy of requiring upgrades before saturation.
For regular participants, I'm not sure that we have a policy.
I'm also not sure if we want a policy there, as that might be considered dictating peering policy. I'm not personally opposed, but this is something that would need some thought.
-- Richard
________________________________
To unsubscribe from the MICE-DISCUSS list, click the following link: http://lists.iphouse.net/cgi-bin/wa?SUBED1=MICE-DISCUSS&A=1
________________________________
To unsubscribe from the MICE-DISCUSS list, click the following link: http://lists.iphouse.net/cgi-bin/wa?SUBED1=MICE-DISCUSS&A=1
-- Jay Hanke, President South Front Networks jayhanke@southfront.io Phone 612-204-0000