DNS RPZ Firewall Rules

Cisco Prime Network Registrar supports RPZ. The DNS firewall rules can be set up for specially designated zones on the Authoritative DNS server. The RPZ and RR data combined with DNS resolver effectively creates a DNS firewall to prevent misuse of the DNS server.

The RPZ firewall rules utilize both the Authoritative DNS and Caching DNS servers to provide the RPZ functionality.

RPZ - Cache DNS

From CPNR 11.2, CDNS queries ADNS for RPZ rule data using zone transfers or notifies and operate on the rule data locally.

The rpz-trigger in the firewall object is removed. Because the data is available locally, this feature greatly increases the performance of processing RPZ data.

Note	The nrcmd> cdns execute expert mode command will have a new option, transfer-zone. It takes the name of an RPZ as parameter. This command causes CDNS to start a probe for any changes to the specified RPZ. For example: nrcmd> cdns execute transfer-zone rpz.example.com

When a query received from client for www.example.com, then CDNS can operate on the rule data locally without querying ADNS.

RPZ - Authoritative DNS

We recommend that you create a separate forward zone on the Authoritative DNS server for RPZ. The zone can be either primary or secondary, and the data can either be manually entered or transferred from a third party RPZ provider. The zones can be named as rpz.<customer-domain> to avoid conflict with domain names in the Global DNS space. In the zone's Query Settings section, enable the rpz attribute to make it an RPZ.

Note	If the RPZ comes via zone transfer, it must be named the same as at the source. If using a commercial RPZ provider, the name is specified by the provider.

Each policy consists of a trigger and an action. The trigger describes when the policy should be applied. The action describes what action should be taken if the policy needs to be applied.

In RPZ zone, each trigger and action combination is defined as a Resource Record (RR). The owner of the RR states the trigger, and the type, and RDATA state the action.

The RPZ RR names can take the following forms:

This zone contains all the RRs related to query names which are in block list. Blocking IP addresses and ranges must be done within the rpz-ip label (that is, rpz-ip.rpz.cisco.com). The same logic can be applied to blocking name servers and clients using the rpz-nsdname, rpz-nsip, and rpz-client-ip labels.

Note	rpz-ip, rpz-nsdname, rpz-nsip, and rpz-client-ip are just another labels and are not real subdomains or separate zones. No delegation points will exist at this level and Caching DNS server relies on finding all the data within the referenced zone.

Note	When using rpz-nsdname and rpz-nsip, the corresponding rule is applied to the original query and will therefore change the answer section. In cases when the final answer is determined from the RPZ rule(s), the RPZ SOA will be included in the authority section.

The Caching DNS server formulates the correct query name, interprets the query response as an RPZ rule, and applies the rule to the client query. If the RPZ rule causes Caching DNS server to rewrite the client response, this data is cached to make future lookups faster. Based on the RPZ zone data, the CDNS server applies the triggers and actions. If no RPZ rule is found, the query proceeds normally.

In addition, RPZ overrides can be configured on the Caching DNS server. This enables the Caching DNS server to override the RPZ action returned by the Authoritative DNS server. This is useful when you do not have control over the Authoritative DNS data as is the case when the data is pulled from a third party. When the Caching DNS server gets a match from the Authoritative DNS server for the RPZ query, it performs the override action rather than the rule action specified in the RR data.

DNS RPZ Actions

RPZ rules are created using standard DNS RRs, mostly CNAME RRs. However, for redirecting, you can use any type of RR. The RR name follows the format based on the RPZ trigger as described in the Table 1 section. The rdata defines the rule action to be taken. The following table describes the RPZ actions.

DNS RPZ Requirements and Best Practices

• All RPZs must have the rpz attribute enabled. A DNS reload is necessary for this change to take effect.

• Both Cisco Prime Network Registrar Authoritative DNS and Caching DNS must be used for end to end RPZ solutions.

• The restrict-query-acl on the RPZ must include only the Caching DNS address and localhost.

• Zone transfers (restrict-xfer-acl) must be either completely denied or restricted only to a specific set of servers.

• RPZ must not be delegated from the parent zone. It must be hidden and only available to a specially configured Caching DNS.

• There must be no RPZ nameserver address record to avoid caching and keeping the name server.

• The name server record must point to "localhost".

• The number of RPZ Firewall rules on a Caching DNS server should be limited to 2-3. The time to process a query increases linearly for each RPZ Firewall rule specified.

• The default TTL, for manually created RPZs, must reflect the rate of change in the zone data. The recommended rate ranges from 5m to 2h.

• The Caching DNS server must revise its max-cache-ttl setting to assure that the cached information is from a reliable source and can be trusted. This setting should be in line with the default TTL of 5m to 2h.

• The Authoritative DNS servers must enable NOTIFY, IXFR, AXFR, and TSIG for zone transfers of the distributed RPZ data.

• An RPZ may contain data for domains (allowed list or block list), but can also be separated into two distinct zones. This can be helpful when there is overlapping data or the block list zone is maintained by a third party (that is, RPZ subscription).