3 - PROVANDV DVS DS 5.5B POST AUDIT REPORT.PDF

Field Audit Report
Dominion Voting Systems Democracy
Suite (D-Suite) 5.5-B Voting System
Maricopa Post-Election Field Audit
Approved by: _________________________________________
Jack Cobb, Laboratory Director
February 23, 2021
v. TR-01-03-MAR-01.00

1.0 INTRODUCTION
The purpose of this report is to document the procedures that Pro V&V, Inc. followed to perform a Post-
Election Field Audit of the Dominion Voting Systems Democracy Suite (D-Suite) 5.5-B Voting System
Maricopa County Board Elections. The Post Election Field Audit was conducted in Maricopa County,
Arizona, from February 2, 2021 through February 5, 2021. The audit was conducted at the following
location:
Maricopa County Elections
510 South 3rd Avenue
Phoenix, Arizona 85003
1.1 References
The documents listed below were utilized in the development of this Report:
 Pro V&V Test Plan No. TP v. 01-03-MAR-01.03, “Dominion Voting Systems D-Suite 5.5-B Voting
System Maricopa Post-Election Field Audit”
 Election Assistance Commission (EAC) 2005 Voluntary Voting System Guidelines (VVSG) Version
1.0, Volume I, “Voting System Performance Guidelines”, and Volume II, “National Certification
Testing Guidelines”
 Election Assistance Commission Testing and Certification Program Manual, Version 2.0
 Election Assistance Commission Voting System Test Laboratory Program Manual, Version 2.0
 National Voluntary Laboratory Accreditation Program NIST Handbook 150-2016, “NVLAP
Procedures and General Requirements (NIST Handbook 150)”, dated July 2016
 National Voluntary Laboratory Accreditation Program NIST Handbook 150-22, 2008 Edition,
“Voting System Testing (NIST Handbook 150-22)”, dated May 2008
 United States 107th Congress Help America Vote Act (HAVA) of 2002 (Public Law 107-252), dated
October 2002
 Pro V&V, Inc. Quality Assurance Manual, Version 7.0
 EAC Requests for Interpretation (RFI) (listed on www.eac.gov)
 EAC Notices of Clarification (NOC) (listed on www.eac.gov)
2 | P a g e
v. TR-01-03-MAR-01.00

1.2 Terms and Abbreviations
The terms and abbreviations applicable to the development of this Test Report are listed below:
“EAC” – United States Election Assistance Commission
“EMS” – Election Management System
“HAVA” – Help America Vote Act
“ICC” – ImageCast Central
“ICP2” – ImageCast Precinct 2
“ISO” – International Organization for Standardization
“NOC” – Notice of Clarification
“QA” – Quality Assurance
“RFI” – Request for Interpretation
“VSTL” – Voting System Test Laboratory
“VVSG” – Voluntary Voting System Guidelines
1.3 Background
The Maricopa County Board of Elections contracted with Pro V&V to conduct a Post-Election Field
Audit to ensure the software and hardware certified for use in Maricopa County are the same as the
software and hardware used in the conduction of the November 2020 General Election. Maricopa also
requested that Pro V&V perform a network analysis and an accuracy test.
1.4 System Description
The D-Suite 5.5-B Voting System is a paper-based optical scan voting system consisting of the following
major components: The Election Management System (EMS), the ImageCast Central (ICC), and the
ImageCast Precinct 2 (ICP2). The D-Suite 5.5-B Voting System configuration is a modification from the
EAC approved D-Suite 5.5 system configuration.
1.5 Scope
The Post-Election Field Audit evaluated the EMS and ICC workstations and servers by comparing the
SHA-256 hash value to the known SHA-256 hash values. In addition, a malware detection tool was run
on each workstation/server to establish whether any malware/virus or malicious software was running on
the workstations/servers. Pro V&V utilized the tool to extract the firmware from a sample of thirty-five
3 | P a g e
v. TR-01-03-MAR-01.00

ICP2 units. These extractions were then placed on the Pro V&V laptop to generate the SHA-256 hash
value for the firmware. These hash values were compared to known hash values for the Election
Assistance Commission Federal Test Campaign. In addition to these evaluations, Pro V&V conducted a
network analysis to ensure the network is a “Closed Network” incapable of reaching the internet Pro
V&V also conducted an Accuracy Test to meet the requirements of the 2005 Voluntary Voting Systems
Guidelines (VVSG).
2.0 AUDIT OVERVIEW
The evaluation of the D-Suite 5.5-B Voting System consisted of removing a copy of the
software/firmware from each component and evaluating the software/firmware against a known SHA-256
hash value outside of the system, running the malware detection tool to verify no malicious software was
resident on the workstations/servers, performing a network analysis, and executing an accuracy test.
3.0 AUDIT PROCESS AND RESULTS
The following procedure outlines the steps that the evaluation team will execute to evaluate the D-Suited
5.5-B under the scope defined in Section 1.5.
3.1 General Information
The evaluation was conducted under the guidance of Pro V&V by personnel verified by Pro V&V to be
qualified to perform the evaluation.
3.2 Audit Configuration
The evaluation utilized system configurations of the D-Suite 5.5-B Voting System and its
components that were setup by Maricopa personnel. Pro V&V had complete access and control
of the equipment being audited.
3.3 Procedures and Summary Findings
ICP2 Software Verification
To perform the verification, the Pro V&V test team randomly selected thirty-five units for
evaluation. A team member then photographed the seals and the device. All seals that needed to
be removed were then removed. After all photographs were taken, the team member removed
any compact flash cards under county supervision and placed them on top of the machine being
evaluated. The team member then inserted two compact flash cards (one blank and the other
containing the firmware extraction tool). The unit was plugged in and powered on with the
security token iButton press on the iButton reader. A password was entered and a tech iButton
was then read by the ICP2 and the option to “Extract Firmware” was selected. The original
compact flash cards were then reinserted into the ICP2. The team member then took the compact
4 | P a g e
v. TR-01-03-MAR-01.00

flash card containing the exported firmware to a Pro V&V laptop to compare the SHA-256 hash
values to the known value from previous testing.
Summary Findings
All SHA-256 hash values retrieved from the units sampled matched the known value from
certification testing. No discrepancies were noted at any time during this portion of the
evaluation.
The serial numbers of the units selected along with the corresponding seal numbers are detailed
in the table below.
Table 3-1 ICP2 Software Verification Serial and Seal Numbers
Seal Number
ICP2 Serial Number
Front Back
FAL19460086 IS143365 1004649
FAL19460030 IS437104 1004719
FAL19330163 IS439376 1004217
FAL19450094 IS419918 1004579
FAL19380033 IS439358 1004621
FAL19460025 IS136178 1004786
FAL19450035 IS441937 1004032
FAL19390009 IS149173 1004260
FAL19380263 IS129272 1004955
FAL19283163 IS1642553 1004904
FAL19450002 IS136177 1004743
FAL19460023 IS437315 1004568
FAL19450257 IS439331 1004216
FAL19320179 IS437217 10041912
FAL19450000 IS1642634 1004973
FAL19450119 ISIS146739 1004997
FAL19252973 IS1642766 1004971
FAL19450133 IS1640855 1004830
FAL19450196 IS1640979 1004572
FAL19380044 IS148896 1004314
5 | P a g e
v. TR-01-03-MAR-01.00

Table 3-1 ICP2 Software Verification Serial and Seal Numbers (continued)
Seal Number
ICP2 Serial Number
Front Back
FAL19460080 IS439339 1004320
FAL19320062 IS439396 1004204
FAL19450068 IS1640786 1004530
FAL19450007 IS1639766 1004747
FAL19450040 IS149919 1004461
FAL19450274 IS439195 1004097
FAL19450241 IS439431 1004375
FAL19460044 IS437295 1004988
FAL19460089 IS437291 1004672
FAL19460042 IS143032 1004752
FAL19450004 IS162418 1004531
FAL19460068 IS437240 1004498
FAL19450034 IS143031 1004491
FAL19450062 IS143686 1004587
FAL19460105 IS1640785 1004125
ICP2 Hardware Verification
To perform the verification, the Pro V&V test team selected five units for evaluation. A team
member then photographed the seals and the device. All seals that needed to be removed were
then removed. After all photographs were taken, the team member removed the necessary
security screws from the bottom of the ICP2. Once the screws were removed the cover was
removed. The team member then used the hardware verification guide to visually inspect the
hardware components and subcomponents against known photographs, part numbers and
identifying marks.
Summary Findings
All units inspected were verified to contain the correct hardware components and
subcomponents. No discrepancies were noted at any time during this portion of the evaluation.
The serial numbers of the units selected along with the corresponding seal numbers are detailed
in the table below.
6 | P a g e
v. TR-01-03-MAR-01.00

Table 3-2 ICP2 Hardaware Verification Serial and Seal Numbers
ICP2 Serial Number Seal Number
FAL19380033 1004580
FAL19450257 ***
FAL19320179 1004481
FAL19320062 1004029
FAL19450040 1004708
***Note: There are various acceptable reasons for a seal to be
unattached, such as: the unit was a spare, the seal was broken
in transit, or the poll worker had to remove it on election night
and return it to the Board of Elections with the elections
results.
EMS and ICC Workstations/Servers Verification
To perform the verification, the Pro V&V test team was granted access to the
workstations/servers from qualified Board of Elections Employees. Once access was achieved, a
team member navigated to the folder containing the DVS software and copied the software onto
a brand new USB drive. The USB was then inserted into the Laboratory laptop and a SHA-256
hash value was generated. A comparison was made between the generated hash value and the
known hash value. The hard drive from the ICC workstation/server was then removed and
placed into a cloning device. The hard drive was then “cloned”. After completion, the hard
drive was place into equipment from Pro V&V’s laboratory that is an exact sample of the same
ICC workstation/server. The equipment was then booted up. The Pro V&V test team was
granted access to the workstations/servers from qualified Board of Elections Employees. Once
that was achieved, a USB containing a malware/virus scanning software was run to scan the
equipment for malware/viruses.
Summary Findings
All units inspected were verified to contain the correct hardware components and
subcomponents. No discrepancies were noted at any time during this portion of the evaluation.
Identification information of the units inspected is detailed in the table below.
7 | P a g e
v. TR-01-03-MAR-01.00

Table 3-3 EMS and ICC Workstations/Servers Verification Details
Computer
Scanner Information
Model Serial Number
ICC Client Workstation
HP-0124K28 OptiPlex 7060 2JGJ3W2
HP-0124K29 OptiPlex 7060 2FDK3W2
HP-0190K29 OptiPlex 7060 2K6M3W2
HP-0192K29 OptiPlex 7060 2JYM3W2
C-GF307234 OptiPlex 3050 8NCCB03
C-GFY00088 OptiPlex 3050 4RMZNX2
C-GF302006 OptiPlex 3050 4RPOPX2
Table 3-3 EMS and ICC Workstations/Servers Verification Details
Computer
Scanner Information
Model Serial Number
C-GFY00019 OptiPlex 3050 4RNZ7X2
C-GFY00347 OptiPlex 3050 4RPVNX2
Adjudicatoin Client Workstation
N/A Dell Precision Tower 3420 87NDHL2
N/A Dell Precision Tower 3431 DVDZG13
N/A Dell Precision Tower 3431 DVFTG13
N/A Dell Precision Tower 3431 G4NFZ23
EMS Client
N/A Dell Precision 3420 27BD8M2
N/A Dell Precision 3420 BNWVCH2
N/A Dell Precision 3420 86PQXK7
N/A Dell Precision 3420 B0ZRMN2
Network analysis
While onsite, qualified Pro V&V personnel evaluated the network architecture to determine the
process and procedure to be followed. All steps were documented in the engineering notebook.
8 | P a g e
v. TR-01-03-MAR-01.00

Summary Findings
Pro V&V test team members evaluated the physical wiring of the network, the managed switch,
clients, and the server. All wiring is housed in an exposed channel hanging from the ceiling.
Different color wires are used for different device types such as printers, PCs, or tabulators. For
the server, commands were run to test connectivity to a known internet address and public IP
addresses. None for these commands returned successful execution from the server or from the
clients. Pro V&V determined that the network it evaluated is a “Closed Network” and does not
have access to the internet.
Accuracy Test
An Accuracy Test was performed to ensure the 5.5-B system correctly captures, stores,
consolidates, and reports the specific ballot selections, and absence of selections, for each ballot
position. To perform the test, the test deck provided by Maricopa Board of Elections was
inserted into each tabulator and processed to reach a total of at least 1,549,703 ballot positions.
Summary Findings
An Accuracy Test was performed on the ICP 2 precinct scanner, ICC HiPro Workstation, and the
ICC Canon DR-G1130 over a two day period. Using the test deck that was provided by
Maricopa County, all votes were tallied and adjudicated resulting in an accurate ballot count. The
ICC workstations were scanned on the first day. Ballots were imported into RTR and
adjudicated resulting in accurate numbers The ICP 2 ballots were scanned on the second day and
were scanned by volunteers from the “Leauge of Women Voters”. Board of Elections staff acted
as poll workers if the volunteers had any issues.
Two anomalies recorded during the execution of this test:
 A ballot jam was recorded on audit unit 10. It could not be determined if the ballot was
tabulated. The Pro V&V test team isolated the ballot until the polls were closed. It was
determined the ballot was tabulated and the ballot was returned to the ballot bin.
 On audit unit 11, after the close of polls it was determined that a ballot jammed and was
rerun through tabulation because the total ballots cast was plus 1. The tabulator was rezeroed
and all ballots were rescanned.
Ballots were imported into RTR and Adjudicated resulting in accurate numbers.
9 | P a g e
v. TR-01-03-MAR-01.00

4.0 CONCLUSIONS
Based on the results obtained during the Field Audit, Pro V&V determines the D-Suite 5.5-B Voting
System, on all evaluated components, is the voting system software and hardware certified for use in
Maricopa County and are the same as the software and hardware used in the conduction of the November
2020 General Election.
10 | P a g e
v. TR-01-03-MAR-01.00