For procurement and engineering teams writing a DPI tender
DPI RFP requirements: a checklist for telecom operators
DPI RFP requirements should state what the platform must classify, where and how it will be deployed, which policy actions and reports it must deliver, how subscriber data is governed, and how bidders will prove their claims on your own traffic before award. This checklist groups the requirements into those areas, lists the evidence to ask bidders for, and ends with an acceptance clause built on a three-stage proof of concept.
A Saudi product, built by Interkey
For procurement, network engineering and regulatory-affairs teams at telecom operators and ISPs. Reviewed November 2026.
- Describe the network first:governance jobs, network points, busy-hour volumes and deployment mode per phase.
- Ask for evidence, not adjectives:per-class results on your traffic, load tests at your peak, and written integration plans.
- Separate out-of-band and inline requirements:classification and reporting do not need the platform in the forwarding path; enforcement does.
- Write the proof of concept into the RFP:acceptance criteria first, then out-of-band classification, then report-only policy before the inline decision.
Part 1: describe your network and the job
Bidders can only size and price a DPI platform against a clear description of where it will run and what it is for. State:
- Governance jobs in scope:capacity planning, policy enforcement, regulatory reporting, content governance for ISPs, security visibility.
- Network points to cover:RAN, core, peering or internet gateways, on fixed networks, mobile networks or both.
- Volumes:busy-hour traffic at each point, and expected growth over the contract term.
- Deployment mode per phase:out-of-band, inline, or out-of-band first.
- Your architecture:the policy and charging architecture and enforcement points, and the reporting and security systems that results must reach.
- Data rules:where data must stay, how long results may be kept, and who may query them.
These are also the factors a DPI quote is built on, so the more precisely they are stated, the more comparable the bids will be.
Part 2: the DPI RFP requirements checklist
Classification
| Requirement | Evidence to ask bidders for |
|---|---|
| Classify traffic by application and protocol at every network point in scope | Detection and false-positive rates for each class, measured on your traffic during the proof of concept |
| Classify encrypted and tunnelled flows, and state whether any decryption is used | A description of the method, and per-class results for the encrypted classes that matter to you |
| Keep classification current as applications change | How models or signatures are updated, and how updates enter your change control |
| Report the share of traffic left unclassified | The unclassified share in proof-of-concept results |
Deployment, scale and resilience
| Requirement | Evidence to ask bidders for |
|---|---|
| Support out-of-band and inline deployment, and state what each mode provides | Architecture for each mode at your network points |
| Handle busy-hour throughput at each point | A load-test plan at your busy-hour peak |
| Defined behaviour if an inline component fails | A failure-mode description and a failover test plan |
| An availability and disaster-recovery design for your topology | The availability design document |
Policy
| Requirement | Evidence to ask bidders for |
|---|---|
| Policies per application, subscriber group or cell, deployed and adjusted in real time | A demonstration during the proof of concept |
| A report-only mode before any rule acts on traffic | A demonstration in report-only mode |
| Support for each policy action on your required list | A written statement, action by action |
| Integration with your policy and charging architecture and 5G core | An integration plan stating what each side provides |
Reporting
| Requirement | Evidence to ask bidders for |
|---|---|
| A network-wide view with drill-down to application and flow | Sample reports and a live demonstration |
| Trends at network and application level, and policy execution feedback | Sample reports |
| Reports in the form your regulatory obligations require | A mapping of each required report to platform output |
| Export to your reporting and security systems | The integration plan |
Data governance, delivery and commercial terms
| Requirement | Evidence to ask bidders for |
|---|---|
| Runs on your infrastructure, with data kept in the Kingdom | Deployment architecture showing where traffic and results stay |
| Output limited to what the governance job needs, with retention and access set by you | A data-flow description listing every data element produced |
| Local delivery: localisation to your traffic mix, integration and support | The support model, team location and working languages |
| A transparent pricing basis | A price breakdown by deployment, capacity, throughput, placement and scope |
Get the RFP checklist as an editable file
The same checklist as this page, ready to paste into your RFP.
Send me the fileLeave your work email below and we send it to you.
Part 3: write the proof of concept into the RFP
Make award or final acceptance depend on results measured on your network, in three stages:
1. Written acceptance criteria.Before anything is installed, agree per-class detection and false-positive thresholds, the reports to be produced, and the expected failure behaviour.
2. Out-of-band classification against verified traffic.The platform classifies your traffic out-of-band, and results are checked against traffic you have verified independently. False positives are measured as carefully as detection.
3. Report-only policy, then the inline decision.Rules run in report-only mode so you see what they would have done. Inline enforcement is decided only after classification has proven itself.
How to set up stage 2 in detail: Measuring DPI classification accuracy on your traffic.
Common gaps in DPI RFPs
- One accuracy figure with no definition.Ask for detection and false-positive rates per class, on your traffic.
- Inline from day one.It puts an unproven platform in the forwarding path. Phase it: out-of-band first, inline when enforcement is needed.
- No false-positive requirement.A false positive means a policy acts on the wrong traffic.
- Reporting obligations left unnamed.List the reports you must produce, so bidders can map them.
- No data-governance clause.State where data stays, what is produced, who may query it and for how long.
- Requirements with no governance purpose.Trace every requirement to a capacity, policy, reporting or compliance need. It keeps the scope, the data produced and the evaluation focused.
How Interkey Intelligent DPI answers these requirements
| Requirement area | Interkey Intelligent DPI |
|---|---|
| Classification | An AI traffic model built into the engine core classifies from flow behaviour; encrypted and tunnelled flows are classified without decryption |
| Keeping current | Automated feature extraction from live traffic, automated model training, and a closed loop from Live View back to models and policies |
| Deployment | Inline or out-of-band on your infrastructure; most operators start out-of-band |
| Scale and resilience | Designed for terabit-per-second traffic intake; full-mesh control and user planes; multi-level disaster recovery; throughput scoped per deployment |
| Policy | Per application, subscriber group or cell; real-time deployment and adjustment; service-chain control; blocking where your policy and applicable regulation require it |
| Reporting | Live View drill-down to application and flow, trend insights, policy execution feedback, latency measured in seconds; a reporting layer for capacity and regulatory teams |
| Integration | Policy and charging and 5G core integration scoped in the deployment design; Interkey integrates with your policy, billing, security and reporting systems |
| Data | Can operate entirely in the Kingdom; output is metadata; retention and access follow your rules and the contract |
| Delivery | Localisation, integration and support from Interkey's team in Riyadh, in Arabic and English |
| Commercial | Quote-based on deployment, capacity, throughput, placement and scope; no published price list |
Questions about DPI RFPs
What should a DPI RFP include?
How is an RFP different from a vendor evaluation?
Should the RFP require inline deployment?
How should accuracy be specified?
Can we discuss requirements with Interkey before issuing the RFP?
Related pages
- How to evaluate DPI vendors
- Measuring DPI classification accuracy
- Inline vs out-of-band DPI
- Network traffic classification engine
- How Interkey runs a DPI proof of concept
- Bandwidth management for ISPs
- DPI in Saudi Arabia
- Policy control and PCRF
- Quality of experience monitoring
- Telecom subscriber analytics
- Tethering detection
- Traffic shaping
- Video optimization
- Zero rating
Proof of concept in three phases
Written acceptance criteria.Agree which rules are tested, on which traffic and subscriber groups, how results are measured and what counts as a pass.
Out-of-band classification against verified traffic.Measure false positives as carefully as detection for every class the rules depend on.
Report-only policy, then the inline decision.Each rule shows what it would have done; enforcement follows only after classification has proven itself.
Discuss your DPI requirements
Tell us the governance jobs in scope, the network points to cover, whether the RFP is in drafting or already issued, and your timeline. Our team in Riyadh replies in Arabic or English.
- Our engineers read your request
- We agree the scope of a technical session
- You see the product on a use case you choose