| 1 |
|
|---|
| 2 |
Intro: |
|---|
| 3 |
|
|---|
| 4 |
The signing tests used two different zones, in two different modes, |
|---|
| 5 |
without opt-in and with. |
|---|
| 6 |
|
|---|
| 7 |
There where 4 different signers used in this test: |
|---|
| 8 |
jdns_signzone, ldns_signzone, pdns_signzone, Bind dnssec-signzone-9.5-pre |
|---|
| 9 |
|
|---|
| 10 |
In this report the names of the actual signers will be obfuscated, |
|---|
| 11 |
signers will be referred to as S1, S2, S3, S4 (not the same order as |
|---|
| 12 |
above). |
|---|
| 13 |
|
|---|
| 14 |
Structure: |
|---|
| 15 |
|
|---|
| 16 |
In the signing test all signers used the same signing parameters |
|---|
| 17 |
Signature Inception time: 20060919000000 |
|---|
| 18 |
Signature Expiration time: 20061231000000 |
|---|
| 19 |
TTL on all RRsets: 3600 |
|---|
| 20 |
NSEC3 iterations: 10 |
|---|
| 21 |
NSEC3 salt: AFFE (hex value) |
|---|
| 22 |
same DNSKEY's 2 keys of alg 5 |
|---|
| 23 |
2 keys of alg 133 |
|---|
| 24 |
|
|---|
| 25 |
By doing this the resulting zones where "directly" comparable, |
|---|
| 26 |
|
|---|
| 27 |
The first zone used [REF1] had 10 different names including |
|---|
| 28 |
apex, there were single label names, multiple label names and a |
|---|
| 29 |
wild-card at multi-label name. There were delegations in the zone, |
|---|
| 30 |
with and without DS records. |
|---|
| 31 |
|
|---|
| 32 |
The second zone was an empty zone (only data at APEX). |
|---|
| 33 |
|
|---|
| 34 |
Expected Results: |
|---|
| 35 |
David Blacka had written a "zone normalizer" to convert the zones to a |
|---|
| 36 |
standard format that then could be compared by diff to find any |
|---|
| 37 |
discrepancies. |
|---|
| 38 |
|
|---|
| 39 |
Actual Results: |
|---|
| 40 |
The normalizer turned out to have two issues in processing the zones, |
|---|
| 41 |
from various signers. The first issue was mixing of cases in names, |
|---|
| 42 |
the second issue was ordering of RRSIG's covering a RRset. |
|---|
| 43 |
Both issues have been addressed. |
|---|
| 44 |
The initial report from the normalizer was that all zones differed. |
|---|
| 45 |
|
|---|
| 46 |
At this point two humans where assigned to figure out the differences |
|---|
| 47 |
in the zones, and what was correct. |
|---|
| 48 |
|
|---|
| 49 |
John's tests |
|---|
| 50 |
Checked the empty zones first. Visual check that all the signed zones |
|---|
| 51 |
had the same |
|---|
| 52 |
number of each type of RR |
|---|
| 53 |
NSEC3-PARAM |
|---|
| 54 |
hashes in the NSEC3 RR's |
|---|
| 55 |
signatures in the RRSIG's |
|---|
| 56 |
TTL's in all RR's |
|---|
| 57 |
|
|---|
| 58 |
Several of the signatures had been done with incorrect |
|---|
| 59 |
parameters. These were resigned. and the checks repeated. |
|---|
| 60 |
Some signers signed zone with both algorithms and some only with alg |
|---|
| 61 |
133. The room determined that this was an open issue that should not |
|---|
| 62 |
be used to judge the signers validity. |
|---|
| 63 |
|
|---|
| 64 |
|
|---|
| 65 |
Olafur's tests: |
|---|
| 66 |
Started work on the non-empty zones, he wrote scripts to check the |
|---|
| 67 |
consistency of the NSEC3 chains and if the NSEC3's matched the |
|---|
| 68 |
NSEC3PARAM record at the apex. |
|---|
| 69 |
The scripts evolved from simply putting out the NSEC3 records to |
|---|
| 70 |
actually checking the chain no matter what order provided. |
|---|
| 71 |
|
|---|
| 72 |
Errors seen: |
|---|
| 73 |
For the empty zones S2 and S1 were found to agree. S4 was found to use |
|---|
| 74 |
the wrong TTL for DNSKEY's and S3 did not create a NSEC3-PARAM RR or |
|---|
| 75 |
its RRSIG. |
|---|
| 76 |
|
|---|
| 77 |
for the non-empty zones: |
|---|
| 78 |
One implementation S4 did not include an NSEC3PARAM. |
|---|
| 79 |
One implementation S1 had the right number of NSEC3 records but |
|---|
| 80 |
multiple NSEC3's pointed to the same target. |
|---|
| 81 |
|
|---|
| 82 |
|
|---|
| 83 |
Second run: |
|---|
| 84 |
After reporting these results the developers sat down and updated |
|---|
| 85 |
their signers to address the issues. |
|---|
| 86 |
|
|---|
| 87 |
After second round of zone signing 3 of the signers (S1, S2, S3) |
|---|
| 88 |
agree on what should be in the NSEC3PARAM record and in the NSEC3 |
|---|
| 89 |
chain. S4 agrees with the the others on the NSEC3 chain but does not |
|---|
| 90 |
include that NSEC3PARAM record in the zone. |
|---|
| 91 |
|
|---|
| 92 |
At this point the opt-out zones where tested and they all passed, all |
|---|
| 93 |
opted out the same name and had the same NSEC3 chain. |
|---|
| 94 |
|
|---|
| 95 |
Conclusions: |
|---|
| 96 |
(any difference between expected and actual results) |
|---|
| 97 |
|
|---|
| 98 |
Signers agree on what NSEC3 records should be for the same NSEC3 |
|---|
| 99 |
settings. All of the signers generate NSEC3 from input flags or signer |
|---|
| 100 |
configuration file. |
|---|
| 101 |
Most of the issues seen where addressed and fixed at the workshop. S4 |
|---|
| 102 |
addressed their issue after testing completed. |
|---|
| 103 |
|
|---|
| 104 |
Open issues: |
|---|
| 105 |
Signing algorithms used. There is an open issue what to do if there are |
|---|
| 106 |
both algorithm 133 and 5 DNSKEY records present. |
|---|
| 107 |
Some singers sign with both algorithms but do NOT include NSEC chain. |
|---|
| 108 |
One signer (S1) does not include the DNSKEY RR's with algorithm 5 when |
|---|
| 109 |
instructed to generate NSEC3 chain. |
|---|
| 110 |
|
|---|
| 111 |
DNSKEY record with algorithm 5 requires RRSIG's on the zone and/or |
|---|
| 112 |
NSEC chain. What does it mean to have NSEC and NSEC3 specific DNSKEY |
|---|
| 113 |
algorithms listed in the DNSKEY set at the apex? |
|---|
| 114 |
|
|---|
| 115 |
Nit: |
|---|
| 116 |
Some singers put their name and version into the signed zone, others |
|---|
| 117 |
did not. Testers found the presence implementation information useful. |
|---|
| 118 |
|
|---|
| 119 |
|
|---|
| 120 |
|
|---|
| 121 |
|
|---|
| 122 |
|
|---|