ToolProtective DNS
DNS security checklist for venues
Score your venue estate against 20 DNS-layer controls across resolvers, policy, logging, bypass, and incident response. You get a readiness score, open gaps, admin hours, and a network policy count. Use the results to set the order of work, starting with bypass routes such as DNS over HTTPS that skip your resolver...
Watch the explainer
Use this to score your venue's DNS security against 20 controls; at 10 controls in place it returns 50 out of 100. The checklist covers resolvers, policy, logging, incident response and bypass, including DNS over HTTPS (RFC 8484). Across 40 locations with three networks each, it returns 10 open gaps, 20 hours of work and 120 policies to map.
What does the DNS security checklist for venues tell you?
The tool returns four figures:
- Readiness score out of 100: how much of the 20-control checklist you have in place.
- Open gaps: the controls still missing.
- Effort estimate: those gaps converted into admin hours.
- Policy count: how many network-level policies you need to map before go-live.
The checklist has five groups of four controls each: resolvers, policy, logging, bypass and incident response. Each group answers one question that an auditor or incident lead will ask:
- Which resolver do your devices use?
- What does it block?
- What does it record?
- Can a device go around it?
- Who acts when it fires?
Where filtering happens, and why it matters
Protective DNS filters at name resolution. A device asks the resolver for a domain before it opens any connection. The resolver checks that domain against threat intelligence and your policy. A blocked domain never resolves, so the device never contacts the phishing page or malware server.
That placement suits venues. You do not manage guest cell phones, a visiting contractor's laptop or the smart TV in a hotel room. You do control what DHCP (Dynamic Host Configuration Protocol) hands out on each network. Point DHCP at a protective resolver and every device that accepts it is covered, with no agent and no new hardware.
The catch is that DNS filtering only protects devices that use the resolver you give them. That is why bypass controls make up a fifth of this checklist.
| Placement | What it inspects | What you need at each venue | Covers guest devices? |
|---|---|---|---|
| Endpoint agent | Traffic on devices you manage | Software on every device | No - guests will not install it |
| Firewall web filter | Traffic through that appliance | A supported firewall at every location | Only where the hardware supports it |
| Protective DNS | Every lookup on the network | A DHCP change per network | Yes, with no agent or appliance |
Flow from a device joining a venue network, through DHCP and the protective resolver, to an allow or block verdict, logging and incident response, with a bypass path stopped at the firewall.
Connections
- Device joins SSID or VLAN to DHCP hands out the Shield resolver
- DHCP hands out the Shield resolver to Shield checks the lookup
- Shield checks the lookup to Allowed: answer returned (clean)
- Shield checks the lookup to Blocked: domain never resolves (flagged)
- Allowed: answer returned to Logs and weekly report
- Blocked: domain never resolves to Logs and weekly report
- Logs and weekly report to Incident response
- Device joins SSID or VLAN to Bypass attempt (ignores DHCP)
- Bypass attempt to Firewall drops it
- Firewall drops it to Shield checks the lookup (lookup returns)
The DNS security checklist and readiness calculator
Work through the checklist with whoever owns your network and firewall configuration. Check a control only when it is live on every network at every location. A control in place at 38 of 40 locations is still a gap, because the guest at location 39 is unprotected.
Download the CSV and record evidence against each line. Useful evidence includes a screenshot, a firewall rule reference or a log export. That evidence turns a self-assessment into something an auditor will accept.
DNS security checklist for venues
0 of 20 done
Count the check marks in each group and enter them below. Then add your venue count, the networks per venue and your own estimate of hours per control.
DNS security readiness calculator
Count the controls in place across all five checklist groups, then enter your estate size.
Readiness score out of 100
50
Open gaps
10
Estimated hours to close the gaps
20
Network policies to map
120
Worked example: With 10 of 20 controls in place, the score is 50, with 10 open gaps and 20 hours of work at two hours per control. At 40 venues with three networks each, you map 120 policies.
Assumptions
- The 20 controls are four from each checklist group: resolvers, policy, logging, bypass and incident response.
- Every control carries equal weight.
- A control counts only when it is live everywhere.
- Two hours per control is a planning placeholder. Replace it with your own figure.
- Every network or VLAN needs its own policy entry.
Reading the four outputs
The readiness score is your headline for a steering group. Open gaps is your work list. Estimated hours tells you whether the work fits this quarter. Network policies tells you how much mapping a resolver change involves. That figure matters most when you scope a deployment or ask for a quote.
Worked example: a 40-site hotel estate
Take an illustrative hotel group with 40 venues. Each venue runs three networks. The first is guest WiFi. The second is Staff WiFi. The third is a back-of-house IoT VLAN (virtual LAN, a logically separate network on shared switches) for door locks, signage, and building management.
The audit found:
- Resolvers, three of four. DHCP and static entries point at the protective resolver, and the service fails closed. Nobody has confirmed DNSSEC validation.
- Policy, three of four. Each network has its own policy, but allow-list changes have no named owner.
- Logging, two of four. Logs show the VLAN and export to the SIEM. There is no retention period and no weekly review.
- Bypass, one of four. Outbound port 53 is blocked. DNS over TLS, DNS over HTTPS, and VPN policy are all open.
- Incident response, one of four. There is an on-call owner and nothing else.
| Checklist group | Controls | In place | Open gaps | Hours at 2 per gap |
|---|---|---|---|---|
| Resolvers | 4 | 3 | 1 | 2 |
| Policy | 4 | 3 | 1 | 2 |
| Logging | 4 | 2 | 2 | 4 |
| Bypass | 4 | 1 | 3 | 6 |
| Incident response | 4 | 1 | 3 | 6 |
| Total | 20 | 10 | 10 | 20 |
The portfolio scores 50. It has 10 open gaps, worth 20 hours at the default rate, and 120 network policies to map.
The score hides where those gaps sit. Six of the 10 are in bypass and incident response. Those are the gaps that matter during an attack. Picture a guest on an infected laptop with DNS over HTTPS (DoH, RFC 8484) switched on in the browser. That laptop goes straight past the policy you spent hours tuning.
Firefox checks a canary domain to decide whether to enable DoH by default. Mozilla documents how networks can signal it. Apple publishes equivalent guidance on preparing your network for iCloud Private Relay. Both belong in your bypass group.
How is the readiness score built?
Every output uses only the inputs you enter:
- Readiness score:
round(100*(resolvers+policy+logging+bypass+incident)/20) - Open gaps:
20-(resolvers+policy+logging+bypass+incident) - Estimated hours: open gaps multiplied by
hours - Network policies:
sites*networks
The assumptions behind each figure
- Equal weights. Each of the 20 controls is worth five points.
- Binary controls. A control half-deployed counts as zero. Partial credit invites optimistic scoring.
- Portfolio-wide hours. The hours input is per control across the whole portfolio, not per site.
- One policy per network. Every network or VLAN gets its own policy entry, even when several reuse one template.
- Identical locations. The calculator treats every location as alike. If yours differ, run it once per location type.
Why the effort figure is a placeholder
Two hours per control is a planning default, not a benchmark. A resolver change on a cloud-managed portfolio can be a single central change. Blocking outbound port 853 may need change approval at every site. Replace the default with your own estimate before the figure goes into a business case.## How do you adapt the checklist to your estate?
You can adapt the tool in three ways: add controls, reweight groups or split by venue type.
Add controls for your sector
When you add a control, raise the divisor from 20 to your new total in the score and gap formulas. Common additions by sector:
- Hospitality and bars: a family-safe category set on guest networks. Staff WiFi gets a separate, lighter policy.
- Travel hubs: a control for passenger networks with high device turnover. Review block reports daily during peak periods.
- Stadiums and events: an incident schedule staffed on event days, when thousands of fans join within an hour.
- Healthcare: separate policies for patients and visitors on guest WiFi and for clinical device VLANs.
- Multi-tenant (Multi-Family, Student Housing, MDU): a documented VPN policy, because residents expect home-like access.
- Retail: point-of-sale devices on their own VLAN with the strictest policy. This keeps shopper traffic out of PCI DSS scope.
Reweight bypass and incident response
To double the weight of bypass, count each bypass control twice and change the divisor to 24. The hotel example then scores 11 of 24, which rounds to 46 rather than 50. The drop reflects the real exposure better than the equal-weight figure.
Split by venue type
A flagship stadium and a branch office rarely share a control set. Run the calculator once per venue type. Then weight the results by the number of venues of each type. The policy count adds up across runs.
Deploying protective DNS per SSID covers DHCP options, resolver mapping and policy per network in detail. Shield applies a policy per network and VLAN. It runs standalone on your existing network or on Purple Access.
What should you do with your score?
Use the score to set the order of work, not as a pass mark.
- Below 50: fix resolvers and bypass first. Without them, every other control is partial.
- 50 to 79: close logging and incident response. These turn blocks into evidence and action.
- 80 and above: test what you have. Try DoH from a test laptop, confirm the attempt appears in logs and time an unblock request end to end.
For the hotel example, the order is bypass (three gaps), then incident response (three), then logging (two). Resolver and policy gaps come last.
Shield meets the third resolver control as delivered: it runs across multiple data centers and fails closed. The DHCP and static DNS changes are yours to make. How Shield works shows what to point where. The Purple platform behind it carries 500 million logins a year at a 99.999% uptime SLA.
DNS controls sit alongside authentication and encryption, not instead of them. The enterprise WiFi security guide covers IEEE 802.1X, WPA3 and segmentation.
Your next step is the logging group. Most estates score lowest there after bypass. Read DNS filtering logs and reporting to decide what to record and who reads it.
Key terms
- Protective DNS
- A recursive resolver that checks each queried domain against threat intelligence and a category policy, and refuses to resolve blocked domains so the device never opens a connection to them. You meet it when choosing where filtering sits on a venue network. It covers unmanaged guest devices with a DHCP change per network, with no agent and no new hardware.
- DHCP
- Dynamic Host Configuration Protocol, specified in RFC 2131, which assigns IP settings to devices joining a network, including the DNS resolver addresses they should use. DHCP is the lever that puts every device on a guest, staff, or IoT network onto the protective resolver. It is the first control in the resolvers group.
- DNSSEC
- DNS Security Extensions, introduced in RFC 4033, which add cryptographic signatures to DNS records so a validating resolver can confirm answers have not been forged or altered. The fourth resolver control asks you to confirm DNSSEC validation for the answers your resolver returns. The 40-site hotel example missed exactly this control.
- DNS over HTTPS (DoH)
- RFC 8484 defines DoH, which carries DNS queries inside encrypted HTTPS traffic on port 443, typically from the browser to a resolver the browser chooses. DoH lets a device skip the resolver your DHCP hands out. Blocking known DoH endpoints is a bypass control and one of the gaps that matters most during an attack.
- DNS over TLS (DoT)
- RFC 7858 defines DoT, which encrypts DNS queries in a TLS session on dedicated port 853 between the client and its chosen resolver. Because DoT uses its own port, you close this bypass route by blocking outbound port 853 at the firewall. That may need change approval at every site.
- Firefox canary domain
- A domain, use-application-dns.net, that Firefox looks up to decide whether to enable DoH by default. A network signals that DoH should stay off by not returning a normal answer for it. Handling the canary domain sits alongside blocking DoH endpoints in the bypass group. Mozilla documents how networks can signal it.
- iCloud Private Relay
- An Apple service that routes Safari traffic and DNS queries through relays, so the local network resolver does not see the lookups. Apple publishes guidance on preparing networks for it. Private Relay, like a VPN, takes DNS out of reach of your resolver. The fourth bypass control asks you to decide and document its policy per network.
- VLAN
- Virtual LAN, a logically separate network on shared switches, tagged under IEEE 802.1Q so traffic on one segment stays isolated from another. Each network and VLAN needs its own named DNS policy, which is why the calculator multiplies venues by networks. In retail, point-of-sale VLANs keep shopper traffic out of PCI DSS scope.
- SIEM
- Security information and event management, a platform that collects, correlates, and retains security logs from multiple sources for detection and investigation. Exporting DNS query logs to your SIEM or ticketing tool is a logging control. It turns blocks into evidence an incident lead can act on.
- Fail closed
- A failure behavior in which a filtering service blocks resolution when it cannot apply policy, rather than passing queries through unfiltered. The third resolver control requires a resolver that runs across multiple data centers and fails closed. Testing failure behavior at least twice a year is an incident response control.
- CCPA/CPRA record of processing
- The record of processing activities required for compliance, listing purposes, data categories, and retention periods for personal data. DNS query logs can reveal browsing behavior, so treat them as personal data. Record the retention period and lawful basis here before go-live.
- PCI DSS requirement 10
- The Payment Card Industry Data Security Standard requirement covering logging and monitoring of access to system components and cardholder data, including log retention. If a network sits inside PCI DSS scope, check its DNS log retention against requirement 10 as well as your CCPA/CPRA retention period.
Worked examples
A hotel group runs 40 venues, each with three networks: guest WiFi, Staff WiFi, and a back-of-house IoT VLAN for door locks, signage, and building management. Where should it start closing DNS security gaps?
The group audited all 20 controls and found 10 in place. That gives a score of 50, 10 open gaps, 20 hours at two hours per gap, and 120 policies to map. The score hid where the gaps sat. Six of the 10 were in bypass and incident response - the gaps that matter during an attack. A guest laptop with DNS over HTTPS switched on in the browser would skip the policy entirely. So the order of work is bypass (three gaps), then incident response (three), then logging (two). Resolver and policy gaps come last.
Equal weighting gives every control five points, so a venue with no bypass controls can still score 60. How do you make the score reflect bypass risk?
Count each bypass control twice and raise the divisor from 20 to 24. Any device using its own encrypted resolver walks past every resolver, policy, and logging control, so bypass deserves extra weight. In the 40-site hotel example, the single bypass control in place becomes two. The estate then scores 11 of 24, which rounds to 46 rather than 50. The lower figure reflects the real exposure better than the equal-weight score. Change the divisor in both the score and gap formulas whenever you add or reweight controls.
Frequently asked questions
How much does Purple Shield cost?
Shield is priced on a tailored quote, scoped to your estate and to whether you run it standalone or on Purple Access. You can run it free for 30 days before you commit. Bring the policy count from the calculator to the conversation. It shows how many networks and VLANs need their own policy, which is the clearest picture of your deployment's size.
Does Shield work with the access points and firewalls we already own?
Yes. Shield works at the DNS layer, so it needs no new hardware. It runs on networks built on Cisco Meraki, HPE Aruba, Ruckus, Juniper Mist, Ubiquiti UniFi, Cambium, Extreme Networks, and Fortinet. You change what DHCP hands out on each network and add firewall rules for the bypass controls. Your existing access points, switches, and firewalls stay in place.
How long does it take to move from another DNS filter?
Migration is a resolver change per network, not a hardware project. Rebuild your category policy in Shield first. Then repoint one network and compare a week of block reports against your current filter. Move the remaining networks once the reports match. Update your bypass firewall rules at the same time, so they allow the new resolver and block the old one.
Do DNS query logs create CCPA/CPRA obligations?
Yes, where a log can be linked to a person or device. Query logs can reveal browsing behavior, so treat them as personal data under CCPA/CPRA. Set a retention period, record the lawful basis and limit who can read them. Purple is ISO 27001 certified and GDPR compliant. If a network sits inside PCI DSS scope, check its log retention against [PCI DSS](https://www.pcisecuritystandards.org/) requirement 10.
How does Shield compare with Cisco Umbrella and Cloudflare Gateway?
All three filter at the DNS layer, so this checklist applies to each. Cisco Umbrella and Cloudflare Gateway sit inside wider secure access platforms built for corporate workforces. Shield is built for venues. It applies a policy per network and VLAN, standalone or on Purple Access, and comes from the team behind WiFi at 80,000+ venues in 90 countries. Compare all three on per-network policy, reporting and failure behavior.
Will DNS filtering slow down guest browsing?
Lookup time is the only point where a DNS filter can add delay, and you can measure it. A lookup happens once per domain, before the connection opens. Shield answers from multiple data centers. Measure lookup times on one network during the trial and compare them with your current resolver. [DNS filtering latency](guide:dns-filtering-latency-speed) sets out what to measure before and after go-live.
Can guests bypass DNS filtering with a VPN?
Yes. A VPN carries DNS inside an encrypted tunnel, out of reach of any network resolver. Decide the policy per network. Block common VPN protocols on Staff WiFi and IoT VLANs, and allow them on guest networks where business travelers need them. Write the decision down. That record is the fourth control in the bypass group.
Is this checklist aligned with NCSC Protective DNS?
It follows the same principle: filter at resolution, keep logs and act on what they show. The NCSC Protective DNS service is for eligible UK public sector bodies, so most venues cannot use it directly. [Our NCSC Protective DNS guide](guide:protective-dns-ncsc-uk-public-sector) covers what venues can borrow from it. The [NCSC's own description](https://www.ncsc.gov.uk/information/pdns) sets out eligibility.
