Choosing, Migrating and Negotiating: A Network Vendor Decision Playbook

The previous five parts covered the market, products, licensing, cost and regional fit. This final part turns that into decisions: how the vendors score, which to choose for which job, how to move off Cisco safely, and how to buy without being caught out at renewal.

Key takeaways

• Cisco (8.10) and Fortinet (8.05) top the weighted scorecard, followed by HPE Aruba (7.85) and Juniper (7.80).
• For a new enterprise network today: HPE or Cisco for campus/WLAN/NAC, Arista for the data center, Fortinet or Palo Alto for security/SD-WAN/SASE — on a standards-based design.
• Negotiate renewals, not just the first purchase: cap escalations, co-term everything and avoid true-forward clauses.

Weighted vendor scorecard

Weights

Scores (1–10)

Scores are relative to full-stack enterprise needs. Security/SASE specialists (Palo Alto, Check Point, Cato, Zscaler) and SP/optical vendors (Nokia, Ciena) score lower on 'technical capability' because they do not cover campus/DC — they may still be the best choice in their category.

Why each vendor received its score

Category leaders

Full-stack ranking (campus + WLAN + WAN + DC)

1. Cisco — most complete and lowest-risk, highest cost.

2. HPE Networking (Aruba + Juniper) — closest functional parity, better AIOps, ~15% lower TCO.

3. Fortinet — best value where security-led branch/campus dominates; weaker DC.

4. Arista — best DC; campus and SD-WAN credible but young.

5. Extreme Networks — simple licensing, education/healthcare strength.

6. Huawei — strong and cheap, but only where regulation and risk appetite allow.

Vendor profiles at a glance

Recommendations by use case and industry

Risk analysis

Higher score = higher risk (1 = very low, 10 = very high). Risk covers vendor lock-in, licensing changes, price increases, supply chain, geopolitics, regulation, security posture, cloud dependency, discontinuation, skills and partner dependency.

If you run Cisco today: what you gain and lose by moving

For each competitor: if an organisation currently uses Cisco, what does it gain and lose by migrating? Migration difficulty: 1 = trivial, 10 = very hard.

Migration playbook

What makes a migration hard

A six-step approach

1. Discover: inventory devices, licences, contract end-dates, proprietary features (EIGRP, TrustSec, VTP).

2. Normalise: migrate EIGRP → OSPF/BGP, PVST+ → MSTP, and SGT → standards-based (VXLAN-GBP / RADIUS VLAN) before swapping hardware.

3. Pilot: one branch + one campus building; validate NAC, wireless roaming, QoS, multicast.

4. Phase by refresh cycle: replace at EoL or contract expiry to avoid paying twice; keep Cisco core during access-layer migration (interop is standards-based).

5. Retrain and automate: budget 5–10% of capex for training and automation (Ansible/Terraform) to realise the opex savings modelled.

6. Exit cleanly: confirm Smart Account licence termination, return-to-vendor credits, and data-wipe certificates.

Typical migration professional-services cost: 10–20% of new hardware value for campus, 15–25% for DC/SD-WAN; duration 3–6 months (campus of 500 users) to 18–36 months (global enterprise).

Procurement playbook

RFP structure

1. Issue a scenario-based RFP (like Scenarios A–D) and require a 5-year TCO with every element priced separately: hardware, perpetual licences, term subscriptions, cloud management, security subscriptions, NAC, SASE, support, optics, services, training.

2. Require list price, discount % and net price per line, per region (INR/EUR/USD), excluding and including GST/VAT.

3. Demand renewal price caps (CPI or ≤5% per year) for all subscriptions and support for Years 2–7.

4. Require co-termination of all subscriptions and an annual true-up (not true-forward) mechanism.

5. Ask for third-party optics support or a fixed optics price list.

6. Evaluate with a weighted scoring model (like the scorecard above) and a hands-on proof of concept (minimum 30 days).

Negotiation levers

• Competitive tension: at least three shortlisted vendors (e.g., Cisco, HPE, Fortinet) per domain.

• Quarter-end/fiscal-year-end timing (Cisco FY ends late July; HPE late October; Fortinet/Palo Alto vary).

• Competitive displacement and Make-in-India/PMA incentives (India).

• Bundle security + networking for platform discounts, but avoid committing beyond realistic device counts.

• Multi-year prepay only with price protection and termination-for-convenience on cloud services.

Top 10 takeaways for network architects and procurement managers

1. The market has consolidated: HPE now owns Juniper (July 2025) and Arista owns VeloCloud (July 2025); re-baseline any vendor shortlists written before mid-2025.

2. Hardware is the smallest part of TCO above branch size — subscriptions, support, SASE/NAC and labour dominate 5-year cost.

3. Cisco remains the lowest-risk choice but typically costs 25–45% more over five years than the cheapest enterprise-grade stack (and ~15–18% more than HPE).

4. HPE Networking is the most complete Cisco alternative; insist on roadmap commitments and price protection during the Aruba–Juniper integration.

5. Arista is the default DC benchmark; pair it with a security vendor because it has no first-party NGFW.

6. Fortinet delivers the lowest TCO for secure branches and SD-WAN, but run a disciplined patch programme for FortiOS edge devices.

7. SASE is converging with SD-WAN; single-vendor SASE (Palo Alto, Fortinet, Cato, Netskope, Cisco, HPE) simplifies operations — test PoP coverage in India and the EU.

8. Huawei is technically strong and cheap, but regulatory exposure (US NDAA, EU 5G Toolbox, India trusted source) makes it unsuitable for most multinational and regulated buyers.

9. Negotiate renewals, not just the first purchase: cap escalations, co-term everything, and avoid true-forward clauses.

10. Design for portability: standards-based protocols (OSPF/BGP, EVPN-VXLAN, RADIUS, gNMI) and automation (Terraform/Ansible) reduce lock-in and make future competition credible.

Planning a network refresh, an SD-WAN decision or a vendor evaluation? I'm happy to talk it through — see consulting or get in touch.

Research date: 8 October 2026. Vendor-neutral and independent — no vendor sponsorship. Prices, licences and product names change frequently; re-validate every figure with at least two authorised partners per region before purchasing. TCO figures are illustrative models built on stated assumptions, not quotations. Gartner content is cited via vendor and press announcements. Not legal, tax or investment advice.

---

The full series: Enterprise Network Vendors 2026

📄 Prefer one document? Download the full 44-page report (PDF)

1. The market after HPE–Juniper and Arista–VeloCloud
2. Products and technical capabilities compared
3. Licensing, pricing and support decoded
4. Five-year TCO across four enterprise scenarios
5. India, Europe and the US: regional realities
6. Choosing, migrating and negotiating (you are here)