Cisco Live AMER 2026: Never underestimate Cisco Live!

Original Entry by : Stephen Donnelly

By Dr Stephen Donnelly, Chief Technical Officer, Endace


Dr Stephen Donnelly, CTO, Endace

Cisco Live AMER is a huge event, attracting more than 20,000 attendees to see the latest products, technology, and innovations from Cisco and its partners. In 2026, Cisco Live AMER had a dedicated SOC for the second time, out on the main show floor for everyone to see. The SOC mission is threefold:

  • To Protect: The primary goal of the SOC is to protect attendees, partners, and staff using the Cisco Live network from threats from outside or within the network
  • To Educate: The SOC educates attendees about modern Agentic SOC operation, current threats, and best practices. Over a dozen new analysts were trained in the SOC, empowered by agentic workflows.
  • To Innovate: The live environment is an opportunity to pressure test how Cisco Security, Splunk Security, and Endace can work together in a next-generation SOC where AI assists, evidence grounds the decision, and humans remain in control

Endace (a Cisco Solutions Plus Partner) was invited to collaborate in the SOC again this year.

We have a great relationship with the SOC team and have assisted at multiple events by providing our always-on full packet capture capability. Capturing all the network traffic at the event not only provides incontrovertible evidence to confirm incident investigations but allows SOC analysts to ‘zoom out’ from a single alert to see the full context of the event.

Visibility of the network activity before and after an incident helps determine the attack chain, and any indications of compromise, such as malware downloads, C2 beaconing, reconnaissance, or lateral movement. A complete record of traffic for the event also enables systematic threat hunting, as well as providing the ability to help train AI on real-world traffic for the Agentic SOC.

The SOC dashboards on show outside the Cisco Live USA 2026 SOC
The SOC dashboards on show outside the Cisco Live AMER 2026 SOC

Endace’s Always-On Packet Capture

To capture all the traffic at the event, Endace supplied two EndaceProbe appliances for use in Cisco’s ‘SOC-in-a-Box’. Each capture appliance features 100 Gigabit Ethernet capture interfaces and NVMe storage, capable of sustained capture at 100Gbps. Each appliance also supports multiple virtual machines which have access to the live captured data in real-time and can host analytics tools. An Endace InvestigationManager VM provides federated search across both appliances, with integrations for Cisco XDR and Splunk, and an MCP server for AI agents.

As incidents are raised in XDR or Splunk Enterprise Security, they are automatically enriched with links to the Endace InvestigationManager. These links allow AI agents and SOC Analysts to easily access full PCAP captures for the related traffic, or pivot to wider searches as needed.

Cisco's SOC-in-a-Box, used at a wide range of events globally.
Cisco’s SOC-in-a-Box, used at a wide range of events globally, is installed in the Cisco Live NOC.

How much is enough?

At 2025 Cisco Live AMER in San Diego, the EndaceProbes captured a total of 78.9TB of network data during the whole event: a record volume of traffic at the time. We expected 2026 to be even bigger, so we configured 160TB of packet storage across the two appliances.

Full packet capture started as soon as the SOC was stood up on the Friday before the show, capturing the network traffic as the show floor and booths were constructed. By the end of the second full day of Cisco Live we had already captured 100TB, and it was looking like our 160TB allocation might be insufficient. This would have meant overwriting traffic captured at the start of the event, so we reconfigured the storage to increase the total packet storage to 192TB. Metadata and search indexes are on top of this figure, allowing retrieval of any captured flow in seconds.

By the end of the show, the EndaceProbe appliances had captured more than 200 billion packets, consuming 199TB of storage, without dropping any packets from the SPAN or overwriting any data since the start of the event. How was that possible with only 192TB of storage configured?

The EndaceProbe appliances support hardware compression of packet data prior to storage. Of the 199TB of traffic collected, 37% was unencrypted, which compresses well. The overall compression ratio for the packet data was 1.25:1, so our 199TB of traffic consumed only about 159TB of storage space. Our original 160TB configuration was pretty close after all!

A dashboard at Cisco Live USA 2026 shows Endace statistics for the show's duraction.
A dashboard at Cisco Live AMER 2026 shows Endace statistics for the show’s duraction.

Why was so much traffic unencrypted, isn’t it a security risk? In some cases yes, the SOC detected 685 unique accounts exposing login credentials in the clear, mostly because of misconfigured email clients using SMTP, IMAP, and POP3 without encryption. The SOC even automated a playbook to contact affected users and warn them of their exposure, inviting them to the SOC for a demonstration and help with remediation.

A surprising amount of traffic still uses unencrypted http. This is commonly used for downloading unencrypted (but signed) software update packages such as Microsoft .cab and Android .apk files, plus EDR rule updates etc. The contents of these files are freely available and frequently cached so transport encryption is not necessary, provided package signatures are verified to confirm authenticity.

The Cisco Live SPAN was configured to capture not only the attendee WiFi network, but also a large amount of infrastructure traffic including CiscoTV, one of the largest users of the network. Some of that traffic was also unencrypted, such as syslog data from IDS appliances. This also contributed to the relatively high percentage of unencrypted traffic seen.

Packet Loss

If the EndaceProbe appliances captured all the traffic from the SPAN without drops and did not overwrite any data, did we get everything? Looking at the bandwidth arriving on the 10GE SPAN port the traffic was reaching 10Gbps early in the day, suggesting the SPAN may be saturated and dropping. Without access to the SPAN port statistics this can be difficult to quantify, but one approach is to look at the sequence numbers seen in TCP flows.

Analyzing traffic flows from Cisco Live USA 2026 in InvestigationManager
Analyzing traffic flows from Cisco Live AMER 2026 in InvestigationManager

When a TCP flow encounters packet loss in the network, a retransmission will occur to fill in the missing data. When this flow is captured by analysis tools they may or may not see the initial lost packet, but they can see the retransmission and infer the loss. Conversely, if the analysis tools see many TCP flows with missing segments but no retransmissions, this can indicate multi-path routing (not applicable at Cisco Live), or packet loss at the SPAN port.

Examining these indicators showed estimated packet loss exceeding 30% during peak times, or around 13-14Gbps of traffic. This could be easily solved in future by upgrading the SPAN port to 25G or 40G Ethernet, a definite learning opportunity.

File Reconstruction

The EndaceProbe appliances also extracted files from unencrypted connections. Over Cisco Live Americas 2026 a total of 2,392,180 files were extracted. These files were filtered by type and deduplicated, with 23,368 submitted to Splunk Attack Analyzer and 5,427 to Cisco Secure Malware Analytics for deeper inspection.

Examining the rate of file extraction over time, we noticed it did not follow the overall bandwidth curve. This is counterintuitive, as more files should be transferred and extracted as the network bandwidth increases during the day.

This anomaly may be caused by the SPAN being overloaded and dropping packets before they reached the EndaceProbes. Files cannot be successfully reconstructed and extracted when packets are lost, and a packet loss rate exceeding 10% implies that most file transfers would have at least one packet missing. With a higher bandwidth SPAN to the SOC we are confident the file extraction counts would have been significantly higher.

Conclusion

Never underestimate Cisco Live! We will be back next year with a faster SPAN and more storage to support the SOC in 2027 (and at Cisco Live APCJ in Melbourne this November), helping protect, educate, and innovate. Hope to see you there!

Acknowledgements

Our thanks go to the Cisco SOC team, led by Jessica Oppenheimer and Ivan Berlinson, for the opportunity to integrate EndaceProbes with the Cisco Live Agentic SOC architecture. 

The SOC team is a collection of Cisco and Splunk experts across many domains who were a pleasure to work and innovate with, and we came away with a great appreciation for the power of the Cisco Security and Splunk tools.

The Endace and Cisco teams were able to prove out integration innovations and test them in earnest in a real-world environment in preparation for making them generally available to the market.

More from Endace, Splunk and Cisco in the SOC series

Read more from Endace about the Cisco Live USA 2026 SOC:

Read related Cisco Team Blogs from the Cisco Live USA 2026 SOC: 
https://blogs.cisco.com/security/clamer-soc-2026

For more Endace blogs in our SOC series, see here:
https://blog.endace.com/tag/soc/ 

Event SOC Website 
Visit Cisco’s Event SOC website for full details of the SOC setup, and download the whitepaper written by Jessica Oppenheimer:
https://www.cisco.com/site/us/en/products/security/event-soc-report.html


Cisco Live AMER 2026: Using LLMs and Endace Full Packet Capture for Incident Response

Original Entry by : Barry "Baz" Shaw

By Barry “Baz” Shaw, Senior Engineering Manager, Technology Partner Programs, Endace


Barry

At Cisco Live AMER 2026 Endace debuted an MCP (Model Context Protocol) server for the integration of EndaceVault into the emergent agentic Security Operations Center (SOC), providing access to 200TB of full packet data and on-demand packet analysis to all agents present in the SOC.

Going into Cisco Live we had key questions about the utility of PCAP data to agents, how to manage making terabytes of data available without overflowing their context windows, and whether the models could parse, analyze and make judgements from the PCAP data comparable to that of an analyst. Results from some simple tests are extremely encouraging and indicate a bright future for agentic AI in the SOC.

Endace Always-on Packet Capture

Two of Endace’s newest EP94C8 EndaceProbes were installed in the GovWare SOC to provide full packet capture to support the conference SOC directives of Protect, Educate, and Innovate.  Full packet capture provides unique insight into all activities on the network, delivering critical context and evidence for Incident Response and Threat Hunting teams.

The EndaceProbes each hosted three VMs that were running Zeek and delivering critical log data into Splunk.  Custom Zeek script additions provided additional valuable details around clear text passwords in email and HTTP, and the use of insecure protocols.  Zeek was also used for file-carving and automated submission of objects to Splunk Attack Analyzer and a beta of Endace’s new Vault API used in a Cisco XDR automated workflow was field tested.   

PCAP and AI – a key question answered

In this scenario we wanted to test the ability of a reasonably new on-prem model to make incident judgements based on PCAP data.  Using the qwen-3.6 model and Open WebUI, a DNS buffer overflow attempt flagged by Cisco Secure Firewall, and raised as an incident in Cisco XDR, was chosen as the candidate.

Cisco Live USA 2026: Possible Buffer Overflow Attempt Flagged
Cisco Live AMER 2026: Possible Buffer Overflow Attempt Flagged

Using the Endace XDR workflow integration (which automatically creates Endace Pivot-to-Vision and EndaceVault links directly in the worklog) a quick pivot from Cisco XDR to EndaceVision shows DNS traffic present:

Investigating the suspect DNS traffic in EndaceVision
Investigating the suspect DNS traffic in EndaceVision

After pivoting from EndaceVision into Wireshark, we can see DNS activity and some large query responses.  This verification served as the analyst judgement to be compared with the LLM findings.

Examining the actual packets in Wireshark lets us validate the AI agent's analysis.
Examining the actual packets in Wireshark lets us validate the AI agent’s analysis.

In Open WebUI a new chat was opened, the Endace MCP connected as a tool, and the following prompt was issued:

There was a firewall alert called ‘PROTOCOL-DNS Microsoft SMTP excessive answer records buffer overflow attempt’ recorded for the traffic between IP addresses <redacted> and <redacted> at 2026-06-03T15:24:01.000Z and 2026-06-03T15:22:33.000Z. Can you get a packet decode from Endace and either confirm or deny the existence of DNS responses that would match this alert.

The model’s reasoning is shown in the following screenshot:

The AI agent's reasoning before starting the analysis.
The AI agent’s reasoning before starting the analysis.

The Endace MCP tools contain detailed annotations that provide context to the model. Above we can see this context helping the model choose between different time inputs and correctly identifying that the start and stop times provided in the prompt were (accidentally) transposed in the prompt.

The AI agent prepares to retrieve and analyze the packet data.
The AI agent prepares to retrieve and analyze the packet data.

Once the AI issued the tool call to the Endace Vault API via the MCP, the extraction of the related PCAP data began. To make this data more consumable by AI, it is converted to JSON by the Endace Vault API and streamed back in manageable sized chunks. The AI can request additional chunks as required, which protects the context window from being swamped by a large PCAP download.

An example of the returned JSONified PCAP data is shown below, this is the raw data that the AI has to make its determination.

A sample of the packet data returned in json format
A sample of the packet data returned in json format

After consuming the JSONified PCAP data the AI began its analysis, clearly having parsed the PCAP data and displaying an understanding of the DNS protocol. From the name of the alert, it understood the specific details pertaining to the alert and was able to look for these in the DNS packets:

The AI agent shows its reasoning and the resulting analysis.
The AI agent shows its reasoning and the resulting analysis – Part 1.
The AI agent completes its analysis and outlines its conclusion.
The AI agent shows its reasoning and the resulting analysis – Part 2.

The reasoning of the model makes for interesting reading. It can comprehend the nature and specifics of the alert and use this to reason about the DNS packets.  It successfully found an oversized DNS response and provided unprompted enrichment by determining the hosting provider of the DNS server.

The AI Agent completes its analysis and outlines its conclusion.
The AI Agent completes its analysis and outlines its conclusion.

So far, the model has performed admirably, validating the firewall alert and the analyst judgement.  To take things further, we asked the model to do some extra analysis over and above the human analysis we’d done. While this chat was user driven, we expect that this is the kind of autonomous activity AI agents will perform on their own.

The AI was asked to confirm the hosting provider and look at the contents of the large DNS response to determine if it contained a malicious payload or was in fact just a verbose and unusually large response.

Can you confirm that the destination host <redacted> is in fact Microsoft hosted?  Is there anything in the large DNS response that could be malicious or is it just a DNS response?

The AI Agent elaborates on its initial conclusion when asked for additional analysis.
The AI Agent elaborates on its initial conclusion when asked for additional analysis.

The model has told us that the source of the DNS response is a reputable domain and that the payload of the response was just a large number of regular AAAA records.  In a pre-AI SOC, the analyst would need to spend time inspecting the PCAP in Wireshark (if they had access to full packet capture in the first place) to determine this was a false positive.

Conclusion

While this example was intentionally not an autonomous agentic finding, the model was able to determine that this alert was a false positive by analyzing the PCAP.  The fact that relatively small open-source models can successfully extract findings from PCAP data is a major validation of their utility in the agentic SOC. We expect that frontier models will provide even richer analysis of PCAP data when available in the SOC if the token cost can be justified. A possible architecture is an on-prem first approach, escalating to a frontier model in the event on-prem cannot make a confident determination.

By using an agent to investigate events like these, and having access to full PCAP data, alerts can be triaged by the agent, requiring only a cursory review by a human analyst to definitively close cases in record time. This validates the core concept that the agentic SOC will improve overall efficiency and accuracy by optimizing incident triage and freeing up time for analysts to work on the more complex incidents that threaten enterprise operations.

Acknowledgements

Our thanks go to the Cisco SOC team, led by Jessica Oppenheimer and Ivan Berlinson, for the opportunity to integrate EndaceProbes with the Cisco Live Agentic SOC architecture. 

The SOC team is a collection of Cisco and Splunk experts across many domains who were a pleasure to work and innovate with, and we came away with a great appreciation for the power of the Cisco Security and Splunk tools.

The Endace and Cisco teams were able to prove out integration innovations and test them in earnest in a real-world environment in preparation for making them generally available to the market.

More from Endace, Splunk and Cisco in the SOC series

Read more from Endace about the Cisco Live USA 2026 SOC:

Read related Cisco Team Blogs from the Cisco Live USA 2026 SOC: 
https://blogs.cisco.com/security/clamer-soc-2026

For more Endace blogs in our SOC series, see here:
https://blog.endace.com/tag/soc/ 

Event SOC Website 
Visit Cisco’s Event SOC website for full details of the SOC setup, and download the whitepaper written by Jessica Oppenheimer:
https://www.cisco.com/site/us/en/products/security/event-soc-report.html