Skip to content
Telecom

Network changes, checked before they run. Against live state and vendor rules.

We check a network change against the live network and the vendor’s rules before the window opens.

Small changes, occasional cascades. Illustrative data.
What is hard here

Reach.

Most network changes go fine. The ones that fail usually touched an element whose current state nobody checked.

What we have learned

One shared element can carry a change far past its cell list, so we trace impact through the topology. After the window, service is compared with its baseline before the ticket closes.

One question in
“Is it safe to push this parameter change to 40 cells tonight?”
What Agent Hive used
  1. sourceChange request CR-5512 · handover offset on 40 cells, cluster 7servicenow
  2. sourceLive state · 3 cells with open alarms, 1 incident in the last 48 halarms, kpi
  3. checkVendor rule · offset range −6 to +6 dB; 2 cells would exceed itguide p.214
  4. checkLicence · feature X2 active on 38 of 40 cellsinventory
  5. workflowImpact · 2 enterprise sites and 1 hospital on first-tier neighbourstopology
  6. resultProceed on 35 cells; hold 5 pending fixes. Sent to the change owner.needs approval
Illustrative data
Figures

Impact is traced.

Fig. 1 · The change runIllustrative data
Live state, vendor rules and impact checked in parallel, then the owner approves.
Fig. 2 · One link changes the reachIllustrative data
In a ring of neighbours, a single shortcut collapses the distance between far cells.
Fig. 3 · Next to the networkIllustrative data
Air-gapped deploymentEverything runs inside the customer’s perimeter with no outbound internet: users, the agent runtime, local models, the knowledge graph, the audit log and read-only connectors to on-prem systems. Updates come in as signed offline bundles. Point at a part to see what it holds and what is blocked.OUTSIDEReleasesigned bundlevX.Y · checksumcarried on offline mediano network linkINTERNEThosted model APIsCUSTOMER PERIMETERNO EGRESSUsersSSO · RBACcustomer identityAGENT HIVE RUNTIMEAgents · workflowsApproval gatesPoliciesEvals · tracesLocal modelsGPU servinginside the siteKnowledge graphdocuments + systemsstored locallyAudit logwho · what · whenevery approvalON-PREM SYSTEMSREAD-ONLY CONNECTORSERPData lakeDocumentsagents read through approved, read-only accessREADUpdate importverify signatureadmin approvesAPPROVED UPGRADEno outbounddata, models, graph and audit never leave the line

Point at a part no egress · signed bundles in · data stays inside

Agents, local models and the audit log run inside the operator perimeter.

Bring us one hard problem. We will say whether it fits our research.