If you haven’t heard by now, there’s a new AI tool called ChatGPT from the OpenAI organization. And if you haven’t tried it, you really should. I myself held back, assuming this AI tool was a lot like my previous exposure to the world of AI, called Midjourney. Midjourney was a neat way for me to kill time and create artwork well beyond my capability. But I got bored with Midjourney after 180 images, and turned back to my regular non-AI driven life. So when my friend told me about ChatGPT, I was highly dismissive. And that was an error.
What can you do with ChatGPT? If you’re an IT person, it is quite eye opening. But one area that IT people run into quite often is diagnosing a common set of symptoms. You can ask it, “why am I seeing X error in Y platform?” And it will answer.
Here’s an example:
ChatGPT Prompt [What I type in]: Why am I seeing giants on my Ethernet interface?
ChatGPT Response: Giants refer to Ethernet frames that are larger than the maximum allowed frame size, usually 1518 bytes. When a switch receives a giant frame, it can cause issues such as:
- Frame drops
- High CPU usage on switches
- Network congestion
Giants can be caused by:
- Incorrect MTU configuration on network devices
- Fragmented frames
- Errors in NIC driver or hardware
- Misconfigured jumbo frames on network devices
To resolve giants, you need to identify the source of the oversized frames and take appropriate measures to fix it, such as adjusting the MTU settings, checking for misconfigured jumbo frames, or upgrading NIC drivers and hardware.
Solid answer from my perspective. I won’t go into the detail, but it is legit.
Let’s try another example. Let’s see if we can get ChatGPT to write a Cisco config.
ChatGPT Prompt: Create an LLQ configuration that prioritizes voice traffic on a Cisco router
ChatGPT Response: To create a Low Latency Queue (LLQ) configuration on a Cisco router that prioritizes voice traffic, you can use the following configuration:
class-map match-all voice
match ip dscp ef
policy-map priority
class voice
priority percent 50
class class-default
fair-queue
interface {interface}
service-policy output priority
You can replace {interface} with the name of the interface on the Cisco router where you want to apply the LLQ configuration.
This configuration creates a class-map named “voice” that matches IP packets with a Differentiated Services Code Point (DSCP) value of 46 (EF), which is commonly used for voice traffic. A policy-map named “priority” is then created that gives the “voice” class a priority of 50% of the available bandwidth and uses fair-queue for the remaining class-default traffic. The service-policy is then applied to the output of the specific interface.
This configuration ensures that voice traffic is always given priority over other traffic, and that remaining traffic is handled fairly.
It’s important to note that, this configuration is based on DSCP marking, for this to work, the voice traffic should be marked with EF at the source and the marking should be preserved through the network.
Additionally, you may also want to consider using other QoS features such as policing, shaping, and congestion management to ensure that your network can handle the traffic and maintain quality of service for your voice traffic.
Again, ChatGPT spits out a config that would make even an experienced Cisco tech sweat.
One thing that I would say about ChatGPT is best in the hands of knowledgeable individuals. Perhaps it is like the One Ring from the Lord of the Rings. In the hand of a mere mortal, it is not to be trifled with. In the hands of someone with real talent, great things can be done. And hopefully through our positive usage of this tool we can counterbalance all those people who use it to cheat on school papers.
