At Cisco Dwell in San Diego, D.J. Sampath, Senior Vice President of Cisco’s AI Software program and Platform group, wowed the group with a demo of AI Canvas. That’s a multi-data, multi-agent system, built-in with Cisco’s AI Assistant and powered by Cisco’s Deep Community Mannequin. In that demo, we may all see AI Canvas’s skill to hurry troubleshooting, carry siloed groups collectively, and allow automation throughout the whole stack.
AI Canvas received’t be accessible till October. Nonetheless, we wished to supply our CCIEs, CCDEs, and Cisco Licensed DevNet Consultants the chance to work with the Deep Community Mannequin as quickly as doable. So we’re making the mannequin accessible to CCIEs and different consultants by means of an AI Studying Assistant accessible in Cisco U.
We predict CCIEs (and shortly, different community engineers) will discover a wealth of ways in which the Deep Community Mannequin may help them study extra and develop into extra environment friendly. However we notice that agentic ops is model new, and that you just may be questioning how one can instantly begin experimenting with the Deep Community Mannequin. So I assumed I’d provide some pattern use instances that can assist you get began.
Tailor-made situations and coaching paths
As a CCIE, you’ve bought years—typically a long time—of expertise in networking, and also you’re totally up to the mark in your group’s IT infrastructure. However what about your staff members, particularly extra junior community engineers? The Deep Community Mannequin AI Assistant can be utilized to construct tailor-made situations and coaching concepts so that everybody in your staff can study the talents wanted for the community you at the moment have, in addition to any new applied sciences your group plans to roll out.
The Deep Community Mannequin understands a variety of networking applied sciences, but it surely’s skilled explicitly on a depth and breadth of Cisco-specific materials. It’s additionally skilled on the supplies and coursework accessible in Cisco U. You may attempt a immediate akin to this one:
- I’m the tech lead for a small staff of community engineers. I must shortly get them up to the mark on the networking know-how we use in our surroundings, together with BGP, MPLS, and OSPF. May you construct me a customized examine plan?
Once I requested this query of the Deep Community Mannequin AI Assistant, I bought a really good syllabus in define type, with hyperlinks to programs in Cisco U.
Right here’s a pattern:
Design validation and optimization
Cisco Validated Designs (CVDs) are primarily blueprints, and IT professionals are accustomed to working by means of them. However typically you want extra steering. The Deep Community Mannequin AI Assistant may help make CVDs extra navigable. It may entry different sources to assist flesh out CVDs and provide recommendations for bettering or optimizing designs.
It may additionally summarize the CVD, supplying you with a high-level overview earlier than studying the entire thing. You possibly can ask it questions akin to:
- Contemplating the CVD for FlexPod, present a getting-started doc that I can use to configure my preliminary UCS supervisor.
- I’m starting to implement the CVD for FlexPod. May you give me a high-level overview of what I’ll be doing and the items I’ll be working with?
The Deep Community Mannequin AI Assistant may help validate an current design with respect to a CVD and provide recommendations for bettering or optimizing designs.
- What sort of storage know-how ought to I think about for booting my blades in a UCS B chassis?
For those who’re having points with a CVD, you may ask the Deep Community Mannequin AI Assistant the place it is best to begin trying.
Automation assistant
The Deep Community Mannequin AI Assistant may assist with automation. You can ask it questions akin to:
- I’m an knowledgeable in community structure and want some assist automating our department SD-WAN deployment. What can be a well-supported, easy-to-learn instrument that may assist me assist this? My staff doesn’t have an excessive amount of coding expertise. May you present examples and hyperlinks to related documentation and coaching?
Troubleshooting
The Deep Community Mannequin AI Assistant may help analyze community diagnostics, akin to syslog messages and debug output, and look at downside signs to offer perception that may be missed by human eyes. Though generative AI continues to be a younger know-how that may make errors, expert-level IT professionals are well-equipped to judge the output for accuracy and detect hallucinations.
For instance, the Deep Community Mannequin AI Assistant may assist interpret a syslog message. You can merely enter the message into the assistant and say you want recommendation or a spot to begin. As a result of it’s skilled on Cisco’s syslog codecs, it may give steering and cross-reference different knowledge.
For those who’re working with a number of knowledge sources, the evaluation turns into extra complicated. With the Deep Community Mannequin AI Assistant, you may describe the symptom you see after which ask, “What ought to I search for?” (In fact, it is best to all the time watch out about pasting uncooked output into AI.) On this approach, you need to use the assistant to information you to the purpose the place you’re snug taking on.
A variety of debugging is basically taking a look at several types of diagnostic knowledge and trying to find the needle in a haystack that can assist you recognize what to do subsequent. The Deep Community Mannequin AI Assistant may help with that course of. For instance, if you should troubleshoot routing adjacencies, you’ll probably want to collect knowledge from a number of gadgets and correlate the information to establish a root trigger.
You are a community troubleshooting assistant. Assist me diagnose why my OSPFv3 session is just not establishing with one neighbor. That is the output from ‘present ospfv3 neighbor’:
OSPFv3 1 address-family ipv6 (router-id 192.0.2.1)
Neighbor ID Pri State Useless Time Interface ID Interface 192.0.2.2 128 EXCHANGE/BDR 00:00:38 13 Vlan300 192.0.2.6 128 FULL/DR 00:00:37 5 Vlan300 And that is the related config from Vlan300: ipv6 deal with FE80::300:241 link-local ipv6 deal with 2001:DB8::241/64 ipv6 allow ipv6 mtu 1500 ipv6 nd dad makes an attempt 0 ipv6 nd ra suppress all no ipv6 redirects ipv6 ospf 1 space 0 bfd interval 1000 min_rx 1000 multiplier 5
Right here’s the response I bought:
In some unspecified time in the future, many people find yourself troubleshooting on the protocol stage (packet seize or it didn’t occur, proper?), the place issues get complicated in a short time. On this case, you may paste the decoded output of a packet seize (akin to that from Wireshark or Tshark) to the Deep Community Mannequin AI Assistant, which may break down the body particulars for you. It may establish hard-to-spot points and dramatically improve the efficacy of deep networking troubleshooting.
The AI assistant may give you extra that means and context than you may get with different instruments. I attempted this with a problematic SNMPv3 packet. The AI assistant appeared on the worth of the fields and defined them to me. Whereas Wireshark confirmed me the sector names, the AI assistant defined that one area, the msgAuthoritativeEngineTime, represented the variety of seconds a tool had been on-line, which was 61411 (roughly seven weeks). The factor is, I simply booted that system. So my SNMP supervisor was confused, and the SNMPv3 entice wasn’t being trusted. Bug discovered!
Whereas most of us are fairly conversant in a variety of community applied sciences, we might not be consultants in each one of many protocols we run on our community. Subsequently, think about how helpful this may be for a protocol you’re not extremely educated about on the area stage. The AI assistant is great at analyzing these fields and explaining their network-relevant context. Whereas the assistant received’t remedy the issue for you, when used correctly, it may give you some good hints. When you perceive extra about these fields, making use of some reasoning and fixing the bug is way simpler.
These are simply a number of the ways in which the Deep Community Mannequin AI Assistant may very well be useful to skilled community engineers. I hope they’re a helpful springboard in your considering. For those who attempt them out, I’d be excited to listen to concerning the outcomes you’re getting.
However I’d be much more excited to listen to about use instances you’ve provide you with that I would by no means consider. AI is an extremely highly effective instrument that may make us extra environment friendly and, frankly, much less confused. However we should determine one of the best methods to make use of them, and we’re all on that journey collectively.
Join Cisco U. | Be a part of the Cisco Studying Community in the present day totally free.
Be taught with Cisco
X | Threads | Fb | LinkedIn | Instagram | YouTube
Use #CiscoU and #CiscoCert to hitch the dialog.
Share: