The GigaPoP receives the vast majority of its Google traffic from MICE. Our regular daytime peak is 20 to 25G, occasionally up to 30ish. On Fri, Aug 23, 2024 at 6:34 AM Miles McCredie <Miles.McCredie@midco.com> wrote:
Curious how many MICE members are in each of the following buckets.
1. Receive limited portion of Google traffic via MICE
Midco and, I assume, US Internet are in this bucket
2. Receive a significant portion of Google traffic via MICE today
I’ve only got a couple years of experience with Google for Midco but my experience has been that Google will not always provide the cache capacity that I think is appropriate. The last two augments were triggered 9-12 months later than I would have liked. Could be my fault for creating new GNLs but…
*Miles McCredie * Principal Network Engineer II-Core IP
Office: 6052755192 Miles.McCredie@midco.com
*Midco.com*
Let's go beyond.
*From:* MICE Discuss <MICE-DISCUSS@LISTS.IPHOUSE.NET> *On Behalf Of *Justin Krejci *Sent:* Thursday, August 22, 2024 5:25 PM *To:* MICE-DISCUSS@LISTS.IPHOUSE.NET *Subject:* [EXTERNAL] - Re: [MICE-DISCUSS] Peering with Google no longer available at MICE IX
*CAUTION:* This email originated from outside of *MIDCO*. *Do not click links or open attachments unless you recognize the sender and know the content is safe.*
I am not sure what Google's stance is on deploying cache servers at an IX vs remote connection like the ones pending disconnection.
If they were to deploy caching nodes at an IX, I assume they would still have a cache-fill requirement in terms of connectivity that has to come from somewhere... of which they appear to be disinterested in maintaining long distance connections going forward. So perhaps this is a donation opportunity for some to entice Google to deploy caching nodes at MICE? Donate local colo and transit bandwidth along with local MICE connectivity?
We have numerous Google caching nodes in our network as well as multiple PNIs to Google, so their pending disconnect from MICE will likely have minimal direct impact on our eyeballs.
-----Original Message-----
*From*: Aaron Wendel <aaron@WHOLESALEINTERNET.NET <Aaron%20Wendel%20%3caaron@WHOLESALEINTERNET.NET%3e>>
*Reply-To*: MICE Discuss <MICE-DISCUSS@LISTS.IPHOUSE.NET <MICE%20Discuss%20%3cMICE-DISCUSS@LISTS.IPHOUSE.NET%3e>>
*To*: MICE-DISCUSS@LISTS.IPHOUSE.NET
*Subject*: Re: [MICE-DISCUSS] Peering with Google no longer available at MICE IX
*Date*: Thu, 22 Aug 2024 22:03:01 +0000
Has MICE requested a cache box from Google?
------ Original Message ------
From "Anthony Anderberg" <AnthonyAnderberg@NUVERA.NET>
To MICE-DISCUSS@LISTS.IPHOUSE.NET
Date 8/22/2024 4:28:18 PM
Subject Re: [MICE-DISCUSS] Peering with Google no longer available at
MICE IX
I've been surprised that folks haven't been talking about this more and am curious about everyone's thoughts - as MICE user-members but also in terms of MICE leadership and direction.
As a user-member I feel somewhat conflicted: on one hand in an effort to maximize performance I could obtain transport to Chicago and connect to Google there, but I feel like doing so is making Google's problem into my problem and that doing so undercuts the goals of MICE. We've seen content providers move MICE's larger members toward local direct private peering, and I can understand rational behind that even if it shifts traffic away from MICE - but Google is proposing remote direct peering at member's cost which feels different.
Of course nobody expects Google to subsidize our corner of the world, I am just curious if there are other aspects we should be thinking about.
Thanks in advance,
Anthony
On 8/22/24, 7:24 AM, "Google Peering Operations No-reply" < peering-ops-noreply@google.com <mailto:peering-ops-noreply@google.com>> wrote:
Dear Peer,
This is a reminder that peering with Google will no longer be available at
MICE Internet Exchange from 16th of September.
As a result of unsustainable overheads associated with connecting to an IX
using remote waves, we have made a decision to stop connecting to IXs that
way. MICE IX falls under the category of IXs we remotely connect to.
We will withdraw prefixes on 16th of September, and disconnect from the IX
on 24th of September.
We realize this may cause inconvenience for some peers. If there is another
mutually present IX that we can peer in, we will be happy to help set up a
bilateral session there - no multilaterals. If there is no mutual IX,
traffic may have to be exchanged via indirect paths (transit links).
We are not looking to connect to any additional IXs to replace the IX that
we are leaving, but if you are considering connecting to another IX and not
sure if Google will also be leaving that IX because of this measure, please
reach out to us and we can share more.
Thank you for your understanding.
Sincerely,
Google Network Operations
------------------------------
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
-- =============================================== David Farmer Email:farmer@umn.edu Networking & Telecommunication Services Office of Information Technology University of Minnesota 2218 University Ave SE Phone: 612-626-0815 Minneapolis, MN 55414-3029 Cell: 612-812-9952 ===============================================