Global AS3356 (Century Link / Level3) Outages Superthread Networking |
- Global AS3356 (Century Link / Level3) Outages Superthread
- META: I guess major news-worthy outages are off topic here?
- Global AS3356 (Level3) Outages
- Global CenturyLink/Level3 Internet Outage (Troubleshooting)
- How do network switches play a role in cluster computing for the forerunner of one extremely power computer? How is latency kept low for the operations?
- Web Proxy | Appliance
- Suggestions for RFCs to read?
- NAT64 Question
- WiFi L3 Roaming
- MPLS question. Two phisycal links between directly connected LSRs.
- Network design/topology to deploy for clients who requires simple flat network that's reliable and cheap
- Assigning a /32 to subscribers.
- Is there such thing as a battery-powered switch/hub?
- Question about some weird DNS
| Global AS3356 (Century Link / Level3) Outages Superthread Posted: 30 Aug 2020 07:32 AM PDT This post is to discuss the Century Link / Level 3 outages. All other threads will be removed. [link] [comments]  |
| META: I guess major news-worthy outages are off topic here? Posted: 30 Aug 2020 06:39 AM PDT Mods, removing the CenturyLink outage thread was an idiotic move. How can you possibly think it was not appropriate in this subreddit, where it targets network engineers? https://www.reddit.com/r/networking/comments/ijb8tn/global_as3356_level3_outages/ Best remove the "news" portion of the subreddit's title if you really feel this thread should have been removed [link] [comments]  |
| Global AS3356 (Level3) Outages Posted: 30 Aug 2020 04:28 AM PDT It looks like some serious routing issues are afoot in AS3356 (CenturyLink/L3). Reports of routing loops in Europe and US. Fastly and Cloudflare seeing impacts: https://www.cloudflarestatus.com/ Anyone been hit by this? [link] [comments]  |
| Global CenturyLink/Level3 Internet Outage (Troubleshooting) Posted: 30 Aug 2020 07:03 AM PDT Could anyone recommend any successful troubleshooting for scenarios like this? I've tried withdrawing my routes, prepending my routes, no-export etc and nothing speeds to be propagating when looking at various route-servers online. This seems to be affecting tons of other networks as well that peer with AS3356, just a glance on DownDetector shows how widely affected the internet at large is. [link] [comments]  |
| Posted: 30 Aug 2020 03:08 PM PDT The part I get is that Ethernet is the most common and most versatile cable used in cluster computing. All cluster computers (from improvised Raspberry Pi clusters to Top 500 supercomputers) use network switches. With the switch, how do operations (calculations, storage, RAM usage) involving the worker nodes interact with other nodes and the master node? If the master node delegates the operations to the worker nodes, how is it done in a way that ensures a continuous equivalent of one extremely powerful computer? What steps are taken to keep latency low? This is to ensure more efficient calculations and other operations. What's the typical and lowest possible latency for cluster computing involving Ethernet? A user on another subreddit recommended asking my questions here. [link] [comments]  |
| Posted: 30 Aug 2020 03:01 PM PDT Hi Guys, Looking for recommendations on a decent proxy, preferably hardware appliance. Looking to serve about 200 persons but should have the ability to scale up to 300. As affordable as possible, even better if it includes content category filtering. Thanks in advance [link] [comments]  |
| Posted: 29 Aug 2020 09:00 PM PDT Yeah, I'm of those nerdy guys who will just read RFCs for the hell of it. I'm looking for suggestions for some RFCs to read. Preferably ones that are conducive to just reading, rather than using as a reference, etc. Any suggestions? What are your favorite? Edit: I should have known better... Please, no April fools RFCs (my favorite is those is hyper text coffee pot protocol) [link] [comments]  |
| Posted: 30 Aug 2020 05:36 AM PDT This is a NOOB question as I have yet to deploy NAT64 and have only began look at it recently. That said, I've heard it referenced as a solution for ISPs that are running out of IPv4 address and want to transition to IPv6. My question is: how does an ISP identify users referenced in DMCA requests or subpoena. Suppose there is a lawful order to reveal who is behind and IPv4 address, but it's only a front for many IPv6 users who NAT to that single IPv4 address. I realize packet inspection would be an option in some cases, but, as far as I know, DMCA strikes are issued off torrents via the public facing IP — which would only show this single IPv4 address. [link] [comments]  |
| Posted: 30 Aug 2020 08:59 AM PDT Does anyone actually use this? Something keeps telling me that it seems like a bad idea but I keep seeing it in different vendor's APs. What is the use case? Why wouldn't I use dynamic VLAN assignment and make every VLAN I require available on all APs? I am not a network engineer. [link] [comments]  |
| MPLS question. Two phisycal links between directly connected LSRs. Posted: 30 Aug 2020 01:27 AM PDT Hi, everyone! There is MPLS Backbone with full-mesh RSVP LSPs between 4 routers(R1 - R4). Please, see the picture below. I've added new phisycal link between R3 and R4(RED curved line). My question: is it possible to include this link into MPLS BB as separate LSP? My goal is to direct specific traffic only in this new LSP and don't let another traffic to go through this. And more general question: is it even possible to have two separate LDP/RSVP sessions between two directly connected(via Ethernet) LSRs? network diagram pic: [link] [comments]  |
| Posted: 30 Aug 2020 06:35 AM PDT I am not "certified" or anything officially. But I am from a computer science background, with a Bachelor's Degree in Computer Applications (doing my master's now). I am confident to work MikroTik's RouterOS for routing/policy based routing, load balancing, firewalling etc to ensure reliable 24/7 connectivity. So basically, I have some potential clients whose requirements are simple:
MikroTik's ultra-cheap routers comes with only 16 MB of flash and a lot of users on their forums have had issues with that. I have been thinking of sticking an RB450Gx4 to all of the clients as it has ample flash and storage + the arm CPU is as good as it gets with long-term software support. The switch/Wi-Fi mesh/APs can be compromised in regards to their "We want cheap" requirements, but I feel I shouldn't compromise on a decent capable router. What do you guys think? [link] [comments]  |
| Assigning a /32 to subscribers. Posted: 29 Aug 2020 04:36 PM PDT With public IPv4 becoming quite scarce now, I have an impetus to save addressing wherever possible. I'm already assigning /31s to customers, but I figure if I could assign /32, I could double the number of subs for my given addressing. I'm thinking that this would be achieved by assigning /32 to the client via DHCP with a gateway not on the same subnet, then proxy-arping or thereabouts for the gateway address. Does anyone do this currently? I've considered PPP, but for resource reasons I prefer not to use it. [link] [comments]  |
| Is there such thing as a battery-powered switch/hub? Posted: 29 Aug 2020 06:31 PM PDT Sorry if there's a better r/ for this. I work on copiers and printers. Sometimes outlets are a premium and extra cords in the case suck. Anyone heard of a battery powered 5 port switch or hub for on the go service? Thanks in advance. [link] [comments]  |
| Posted: 29 Aug 2020 05:10 PM PDT I work in software support, networking knowledge is pretty minimal -- basically just enough to know when/how users need to enter their creds to get to the server's file share and when there's an issue. So an office called with an issue the other day -- software can't connect to the SQL database. Usually this is because they just booted the database computer up and since we set the SQL service to delayed start, it takes upwards of 5 minutes for the database to be available. Not the case this time, instead I could navigate to the share just fine in File Explorer, but when I ran a ping, the IP came back 142.xxx... something super weird. Even stranger? when I ran I changed the connection parameters in the software to just use IP after I verified it was static, but being software support, all I could do was tell them to bring the issue up with their netadmins. But I'm a learner and would really like to know what causes this, as I've seen it a couple times. [link] [comments]  |
| You are subscribed to email updates from Enterprise Networking Design, Support, and Discussion.. To stop receiving these emails, you may unsubscribe now. | Email delivery powered by Google |
| Google, 1600 Amphitheatre Parkway, Mountain View, CA 94043, United States | |
No comments:
Post a Comment