❄ J ❄

2K posts

❄ J ❄

❄ J ❄

@jbemmel

➀+➊=♵ Mostly flawed human with a hint of reality

Houston, TX เข้าร่วม Temmuz 2010
78 กำลังติดตาม216 ผู้ติดตาม
ทวีตที่ปักหมุด
❄ J ❄
❄ J ❄@jbemmel·
❄ J ❄ tweet media
ZXX
0
0
0
104
❄ J ❄
❄ J ❄@jbemmel·
What if the universe is fundamentally hyperbolic, and redshift reflects an accumulated twist along the path?
❄ J ❄ tweet media
English
0
0
0
32
❄ J ❄
❄ J ❄@jbemmel·
@BestBuy please fix your order pickup process - just spent an hour in your store waiting for the system to release my online purchase to me A $1000 Black Friday deal became a huge waste of time - not worth it
English
2
0
0
24
❄ J ❄
❄ J ❄@jbemmel·
@jwbensley I imagine the risk/reward calculation for most network operators would tilt towards not doing that in production. Researchers and academics - fun challenge
English
0
0
0
21
alexis
alexis@digitalbyte_·
when I first learned about the undersea cables, I thought my instructor was pulling my leg. hope y’all think this is as cool as I do 🤓
English
9
9
120
5.7K
Kevin Nanns
Kevin Nanns@AdjacentNode·
How a network engineer keeps his wife happy with a magic button. Wait…
English
6
5
61
5.5K
Tim Bertino
Tim Bertino@TimBertino·
Can anyone give some examples of production network use cases that would require proxy ARP? Just trying to understand what scenarios it would make sense.
English
20
0
19
7.6K
Orhan Ergun
Orhan Ergun@OrhanErgunCCDE·
The best comment wins! Router on a stick.
Orhan Ergun tweet media
English
54
17
140
15.8K
❄ J ❄
❄ J ❄@jbemmel·
@satyanadella s/IT systems/Windows systems/g The industry would do well to diversify and safely switch their systems to Linux, supported by a broad ecosystem of vendors. Microsoft is the single point of failure here
English
0
0
0
59
Satya Nadella
Satya Nadella@satyanadella·
Yesterday, CrowdStrike released an update that began impacting IT systems globally. We are aware of this issue and are working closely with CrowdStrike and across the industry to provide customers technical guidance and support to safely bring their systems back online.
English
2K
5.1K
51.5K
5.6M
❄ J ❄
❄ J ❄@jbemmel·
@mbushong Sorry to hear about your wife, hope she recovers soon
English
0
0
0
71
❄ J ❄
❄ J ❄@jbemmel·
@ntdvps An extensible model-driven API to query state and configure the network using a consistent, unified human friendly interface "Show me the top ten customers in terms of traffic volume" "How much room is left in my route tables?" "Enable Segment Routing on the backbone routers"
English
0
0
0
66
Roman Dodin
Roman Dodin@ntdvps·
A powerful, 100% streaming, state-less and real-time Network Query API is coming to your networks this fall 🍂 Operational visibility gets much better when you have a chance to work on a network-wide level, instead of on a per-device level. With natural language support, of course. Be honest, it is time to do better than just grepping for "oper-state up" in a stream of text over ssh.
English
4
3
35
3K
Roman Dodin
Roman Dodin@ntdvps·
Still feels surreal that one can work on preparing 70 VMs for Autocon1 attendees while being 10k feet above the ground 🧐 Pst, I drained the private cloud, no more 6cpu/24GB VMs left, oopsie. I also went and quickly checked what would it cost me to run 70 8vcpu/32GB VMs in public cloud for 24 hours, and it is not cheap :D
Roman Dodin tweet media
English
6
1
57
5.1K
❄ J ❄
❄ J ❄@jbemmel·
@ntdvps So what’s the next hurdle? ✅ Freely available image ✅ low resources …
English
1
0
0
87
Roman Dodin
Roman Dodin@ntdvps·
And while having an accessible image is a first step to make, the other one is making your image consumable. If a single instances demands 8vcpu/32gb mem and 15 minutes to boot -- you are doomed to have a negative overall experience
Danny Wade@devnetdan

As someone who focuses on infra/network automation, I couldn't agree more. Having stable VM/container images widely available for testing and development is a must. blog.ipspace.net/2024/05/too-st…

English
5
1
28
3.1K
Daniel Dib
Daniel Dib@danieldibswe·
You have two L3 devices. You want to run a routing protocol like OSPF or BGP between them. You need multiple links for load sharing. Would you do: 1. L2 bundle with SVIs. 2. L3 bundle with routed ports. 3. Routed links with ECMP. Motivate your answer.
English
70
15
120
26K
❄ J ❄
❄ J ❄@jbemmel·
@AdjacentNode I think the 16U might be a problem, if you can convert to a 4x4U layout you may have more options
English
0
0
0
13
Kevin Nanns
Kevin Nanns@AdjacentNode·
I have a 16U rack for my network and server gear that’s currently in the garage because the wife doesn’t want it in the house. She conceded that I can bring the lab in the house if I can make it look like furniture. Anyone know of a nice looking solution?
English
57
4
84
17.4K
❄ J ❄
❄ J ❄@jbemmel·
@martin_casado Not just "application logic", all logic and value creation is being abdicated. As in "we're using an LLM (doing 'AI') and that's the application" It's a consequence of abandoning explainability. Now we have a black box that occasionally produces some interesting output
English
0
0
1
29
martin_casado
martin_casado@martin_casado·
One thing that's so unique about current use of LLMs is that application logic is being abdicated to them. I can't think of any other time this has happened. Normally infrastructure moderates resource usage, not logic and correctness.
English
19
13
123
30.2K