Current Architecture Analysis¶
Existing Infrastructure (Confirmed from Documentation)¶
SD-WAN Architecture: DIA Model¶
Transport Strategy:
| Site Type | Primary Transport | Secondary Transport | Internet Breakout |
|---|---|---|---|
| Hub Sites (6 sites) | MPLS | Dedicated Internet | Local DIA via Umbrella |
| Large Branches | MPLS | Internet DIA | Local DIA via Umbrella |
| Medium/Small Branches | Internet DIA | 5G/LTE | Local DIA via Umbrella |
Key Characteristics: - ✅ Local internet breakout at every site - ✅ Umbrella SASE deployed with IPsec tunnels from all WAN edges - ✅ Application-aware routing: Corporate traffic → MPLS, SaaS → DIA - ✅ Cloud-direct paths: No traffic hair-pinning through Mumbai/London/NJ hubs
Traffic Flow (Current):
┌──────────────────────────────────────────────────────────────────────────┐
│ CURRENT DIA TRAFFIC FLOW │
├──────────────────────────────────────────────────────────────────────────┤
│ │
│ Branch User (Delhi) │
│ ┌──────────────┐ │
│ │ Accesses: │ │
│ │ • Office 365 │ │
│ │ • Salesforce │ │
│ │ • YouTube │ │
│ └──────┬───────┘ │
│ │ │
│ ▼ │
│ ┌──────────────────────────────────────────────────────┐ │
│ │ SD-WAN Edge (ISR 4331) │ │
│ │ ──────────────────────────────────────────────────── │ │
│ │ Policy Decision: │ │
│ │ • Corporate (10.252.x.x) → Route via MPLS │ │
│ │ • Office 365 (52.96.x.x) → Route via DIA │ │
│ │ • Salesforce → Route via DIA │ │
│ │ • YouTube → Route via DIA │ │
│ └───────────────────┬──────────────────────────────────┘ │
│ │ │
│ ▼ │
│ ┌──────────────────────────────────────────────────────┐ │
│ │ Umbrella SASE (Mumbai PoP) │ │
│ │ ──────────────────────────────────────────────────── │ │
│ │ • DNS Security: Block malicious domains │ │
│ │ • Web Filtering: Block gambling, adult sites │ │
│ │ • Cloud Firewall: Inspect Layer 4 traffic │ │
│ │ • IPS: Detect/prevent exploits │ │
│ └───────────────────┬──────────────────────────────────┘ │
│ │ │
│ ▼ │
│ ┌──────────────────────────────────────────────────────┐ │
│ │ Internet / SaaS Applications │ │
│ │ ──────────────────────────────────────────────────── │ │
│ │ • Office 365: <50ms latency (direct path) │ │
│ │ • Salesforce: <80ms latency │ │
│ │ • YouTube: Cached at ISP/CDN edge │ │
│ └──────────────────────────────────────────────────────┘ │
│ │
└──────────────────────────────────────────────────────────────────────────┘
WxCC Deployment (Cloud-Based)¶
Existing Configuration:
| Component | Deployment | Details |
|---|---|---|
| Webex Calling | Cisco Cloud | 3,200 users, migrated from on-premise CUCM |
| Webex Contact Center (WxCC) | Cisco Cloud (US Region) | 175 agents, 5,000 calls/day |
| Call Recordings | Cisco Cloud Storage | 90-day retention, compliance |
| Integration | Salesforce CRM | Screen pop, CTI |
| PSTN Gateway | On-Premise CUBE/SBC | Mumbai, London, NJ sites |
Current WxCC Analytics Limitations: - ❌ No real-time sentiment analysis - ❌ No AI-powered agent coaching - ❌ No churn prediction - ❌ No IVR optimization with NLU - ❌ Manual call quality reviews (QA team listens to recordings)
Goal: Integrate GCP Vertex AI to enable advanced analytics