COOP_Proposers_Day_Slides.pdf

Continuous-correctness On Opaque Processors (COOP) Proposers Day

Dr. Lok Yan

March 18th, 2024

Distribution Statement A – Approved for Public Release, Distribution Unlimited

Agenda

Start End Time Speaker

8:00 AM 9:00 AM 60 Self

9:00 AM 9:05 AM 5 DARPA Security

9:05 AM 9:20 AM 15 DARPA Contracting

9:20 AM 9:35 AM 15 DARPA Connect

9:35 AM 10:20 AM 45 Dr. Lok Yan

10:20 AM 10:25 AM 5 Dr. Whitney Mason

10:25 AM 11:00 AM 35 Self

11:00 AM 12:30 PM 90 Self

12:30 PM 1:00 PM 30 Dr. Lok Yan

Self

Self

Dr. Lok Yan

5:00 PM 5:00 PM 0 Self

1:00 PM 5:00 PM 240

Questions & Answers

MTO Overivew

Amentum - Arlington, VA

Monday, March 18, 2024

Breakout - PM Technical One-on-One

Description

Event Check-In

DARPA Contracting Overview

DARPA Connect

Technical Overview

Proposers Released

Lunch | Proposers Released

Breakouts

Security Overview

Open Discussions

Lightning Rounds

Security Overview

DARPA Security

This meeting is:

UNCLASSIFIED

• Laptops and cell phones are NOT authorized within this meeting.

• Photography, video, and audio recordings of any kind are prohibited.

Security

Distribution Statement A – Approved for Public Release, Distribution Unlimited 4

COOP has been designated as an Unclassified Applied Research Program and will likely have CUI.

COOP will use the MTO CUI Guide.

For information on how to handle CUI, see the National Archives information page:

https://www.archives.gov/cui

• CUI is information the Government creates or possesses, or that a contractor creates or possesses for or on behalf of the Government. CUI requires safeguarding or dissemination controls.

• COOP Reports and presentation slides including CUI must be appropriately marked

• A CUI Designation Indicator/Warning box is required on the first page, cover of all CUI containing documents – see next page example

• “CUI” in the header and footer of the document is also required

• An appropriate dissemination control will be placed at the bottom of every page/slide, i.e. “Federal

Employees and Contractors Only (FEDCON)” or “No Dissemination to Contractors” (NOCON)

• Portion mark all section of the document with either (CUI) or (U) for unclassified information not containing CUI

Controlled Unclassified Information (CUI)

Distribution Statement A – Approved for Public Release, Distribution Unlimited 5 https://www.archives.gov/cui

Military Systems and Space Applications

• If information is produced linking military systems or space applications, it is considered Controlled

Technical Information (CTI)

• Be cognizant of your schedule and of mission relevant use cases

Export Control

• Consider export controls and how they impact the program when proposing or suggesting technical solutions

Classified / SAP Submissions

• Proposers are expected to provide unclassified proposals. However, if classified or SAP proposal submission is required contact COOP BAA team prior to submission

• Coordinate with original customer to obtain approval to share information and provide security classification guide (SCG)

Technical information

DARPA contracting overview

Brian Nuckols

A copious amount of information is made available to potential proposers to clarify program goals/objectives and proposal preparation instructions.

However:

• Only the information/instructions in the BAA counts

• Proposals will only be evaluated in accordance with the instructions provided in the BAA

• Any response provided by the Government in the FAQ that’s different than what is provided in the BAA will be made formal by an amendment to the BAA

• Such responses will make note of an impending BAA amendment

Only a duly authorized Contracting Officer may obligate the Government

Proposers Day Disclaimer

1. BAA published: March 13, 2024

2. Proposers Day: March 18, 2024

3. Abstracts Due: March 28, 2024

4. Government Abstract Responses: ~mid April 2024

5. FAQ Submission Deadline: May 03, 2024

6. Proposals Due: May 13, 2024

7. Proposals are checked for BAA compliance

Non-conforming proposals are not reviewed (and cannot be selected)

8. Government conducts Scientific Review Process

Clarification requests may be sent to various proposers

9. Government sends out notification letters: ~June 2024

10. Contracts negotiated & awarded (performance start): ~September 2024

See BAA for complete date/time information Monitor SAM.gov for any BAA amendment(s)

BAA process/timeline

• Described in Federal Acquisition Regulation Part 35:

• For the acquisition of basic and applied research

• FAR 35.016 (a): Shall only be used when “meaningful proposals with varying technical/scientific approaches can be reasonably anticipated.”

• BAAs allow for a variety of technical solutions

• The BAA defines the problem set, the proposer defines the solution (and SOW)

• DARPA Scientific Review Process

• Proposals are evaluated on individual merit and relevance as it relates to the stated research goals/objectives rather than against one another

• Selections will be made to proposers whose proposals are determined to be most advantageous to the

Government, all factors considered, including potential contributions to research program and availability of funding

Government may select for negotiation all, some, one, or none of the proposals received Government may accept proposals in their entirety or select only portions thereof Government may elect to establish portions of proposal as options

BAA overview (1 of 3)

• Multiple Awards Anticipated

• Award instrument types:

• Procurement contract

• Other transaction

• Phase I

• Base Period

• 18 months

• The goal of Phase I is to demonstrate a simulated COOP solution on multi-threaded cores

• Phase II

• Option Period (Option 1)

• 18 months

• The goal of Phase II is to demonstrate the generalizability of the COOP solution on multiple different processor types

BAA overview (2 of 3)

• Supplemental Options

• Two additional options (Options 2 and 3) should be priced for a potential Government-funded chip fabrication and packaging activity

• Option 2

• 6 months

• Detailed design option to bring the COOP solution to tapeout (i.e., GDSII only)

• May be exercised at any time during Phase I

• Option 3

• 7 months

• COOP fabrication and packaging effort.

• May only be exercised if Option 2 had been exercised

• Associate Contractor Agreements

• The success of COOP depends upon the open exchange of information across program performers

• Each resultant award will contain a term/condition classifying the performer as a Associate Contractor and requiring that the performer enter into an Associate Contractor Agreement with all other COOP performers

BAA overview (3 of 3)

• All interested/qualified sources may respond subject to the parameters outlined in BAA

• Foreign participants/resources may participate to the extent allowed by applicable Security Regulations, Export Control Laws, Non-Disclosure Agreements, etc.

• FFRDCs and Government entities:

- Are not prohibited by the BAA from proposing

- Are, however, subject to applicable direct competition limitations

- Are, however, required to demonstrate eligibility (sponsor letter)

The burden to prove eligibility for all such team members rests with the proposer Request that any FFRDC/Government entity contemplating participating as a prime/sub contact DARPA in advance (see BAA)

• Real and/or Perceived Conflicts of Interest:

- Identify any conflicts

- If any are identified, a mitigation plan must be included

Eligibility considerations

• Abstracts are strongly encouraged:

1. They minimize unnecessary effort in proposal preparation and review

2. They reduce the potential expense of preparing an out of scope proposal

• The abstract provides a synopsis of the proposed project (tech and budget)

• Government will reply by letter with one of two possible responses:

1. Encourage full proposal, and may provide feedback

2. Discourage full proposal, and will provide feedback

DARPA will not communicate further (verbally or in writing)

• Regardless of DARPA’s response to an abstract, proposers may submit a full proposal

DARPA will review all conforming full proposals submitted without regard to abstract recommendation/feedback

Abstracts (1 of 2)

• Abstracts must include the following components:

• Executive Summary

• Informal Threat Model and Assumptions

• Technical Approach

• Capabilities/Management Plan

• Bibliography

• Submit UNCLASS submissions to the DARPA BAA Website (https://baa.darpa.mil).

Abstracts (2 of 2)

Max of 4 pages

The total page limit does not include the cover sheet (required), table of contents (optional), bibliography (required), or technical papers (optional, no more than 3) https://baa.darpa.mil/

• Volume I: Technical and Management Proposal

• Be mindful of the page limitation (15 pages (see BAA for exclusions))

• Be sure to respond to all of the required items

• Executive Summary

• Innovative Claims Table

• Informal Threat Model and Assumptions

• Technical Approach

• Management Plan

• Requirements for teams that include Government entities and/or FFRDCs

• Organizational Conflict of Interest affirmations and disclosure

• Capabilities

• Novelty of proposed work

• Schedule and milestones

• Appendix I: Formal Threat Model

• Appendix II: Bibliography (Optional)

• Appendix III: Statement of Work – Organize by Phase/Option, then by Tasks. Define all tasks (what you are doing, not how you are doing it) and deliverables (data, software, and material items, as applicable).

• Appendix IV: Administrative

• Pay attention to key technical elements in the “Funding Opportunity Description” that must be addressed in all proposals

Full proposal preparation (1 of 3)

• Volume II: Cost Proposal

• No page limitations

• Fully detailed cost build-ups/estimates (Prime & Subcontractors)

• Summary Cost Build-up: By phase and performer fiscal year

• Detailed Cost Build-up: By phase, technical task, and month

• Subcontractor proposals are required (to include SOW)

• No Rough Order of Magnitudes (ROMs)

• Certificate of Current Cost/Pricing Data will be required for any procurement contract award meeting the prescribed threshold (i.e., the adequate price competition exemption does not apply to BAAs)

DARPA Standard Cost Proposal Spreadsheet The Government requires that proposers use the provided Attachment E: MS ExcelTM DARPA Standard Cost Proposal Spreadsheet in the development of their cost proposals. All tabs and tables in the cost proposal spreadsheet should be developed in an editable format with calculation formulas intact to allow traceability of the cost proposal. This cost proposal spreadsheet must be used by the prime organization and all subcontractors. In addition to using the cost proposal spreadsheet, the cost proposal still must include all other items required in this announcement that are not covered by the editable spreadsheet. Using the provided cost proposal spreadsheet will assist the Government in a rapid analysis of your proposed costs and, if your proposal is selected for award, speed up the negotiation and award execution process.

Full proposal preparation (2 of 3)

• Fee

• Alternate structured approach per DFARS 215.404-4(b)(1)(C)

• Established fee calculation range for COOP: 6.0% - 8.8%

• Factors considered:

• Performance risk

• Contract type risk

• Facilities capital employed

• Anticipated award size

• Available transition path

• Markets (commercial, Government, international)

• Intellectual property rights

• Chances of award

• Time to production

• Solicitation complexity

• No fee justification necessary in proposal

• No fee negotiation

Elimination of fee as a negotiation item is expected to result in reduced contracting timelines for any proposal selected for award negotiation

Full proposal preparation (3 of 3)

• Government desires as few restrictions as possible

• A minimum of Government Purpose Rights is the expectation

• If asserting less than Unlimited Rights (e.g., Restrictions):

• Provide and justify basis of assertions using the prescribed format

• Explain how each item will be used to support the proposed research project

• Explain how the Government will be able to reach its program goals (including technology transition)

The proposer (prime) must submit a Data Rights Cert covering the entire team (prime and subcontractors), as applicable

Provide a Data Rights Cert no matter the instrument type being proposed This information is assessed during evaluations (barriers to transition)

Data rights

• Failure to submit proposal on time

• Proposal due date and BAA closing date are the same – so, late is late!

• Failure to submit using the correct mechanism

• DO NOT use the DARPA BAA website for this BAA

• All proposal submissions are anticipated to be unclassified

• Failure to submit both proposal volumes

• Pages beyond the page limitation (Vol. I) – additional pages will not be reviewed

• ROM/s instead of full subcontract cost proposal(s)

• “I didn’t have time to get the subcontract proposal(s)” will not change the outcome

• “My subcontractor/s would not give me a proposal” will not change the outcome

Non-Conforming proposal determinations

• Prior to Receipt of Proposals (Solicitation Phase): Follow instructions in BAA Typically handled through the FAQ, but see BAA exceptions

• After Receipt of Proposals/Prior to Selections (Scientific Review Phase): Limited to Contracting Officer or BAA Coordinator (with approval) to address clarifications requested by the review team Proposal cannot be changed in response to clarification requests

• After Selection/Prior to Award (Negotiation Phase): Negotiations are conducted by the Contracting Officer PM and/or COR typically tasked with finalizing the SOW (with PI) PM and/or COR typically involved in any technical discussions (i.e., partial selection discussions) Pre-award costs will not be reimbursed unless a pre-award cost agreement is negotiated prior to award

• Informal Feedback Sessions (Post Selection): May be requested/provided once the selection(s) are made

• If made on a timely basis (~2 wks after letter), all requests will be accepted

Communications

• Proposer Instructions and General Terms and Conditions

• Proposer Instructions: Procurement Contracts

• Proposer Instructions: Other Transactions

• https://acquisitioninnovation.darpa.mil/

• www.DARPAConnect.us

Helpful links https://www.darpa.mil/work-with-us/proposer-instructions https://www.darpa.mil/work-with-us/procurement-contracts https://www.darpa.mil/work-with-us/other-transaction-agreements https://acquisitioninnovation.darpa.mil/ http://www.darpaconnect.us/

DARPA Connect

Sana Hoda

COOP Proposers Day Microsystems Technology Office (MTO) March 18, 2024

Distribution Statement A – Approved for Public Release, Distribution Unlimited 24

Regional and Virtual Events

DARPAConnect is designed to broaden DARPA’s reach and stimulate growth and collaboration between DARPA, businesses, and academia.

Breaking Down Barriers to Entry for Nontraditional Performers www.DARPAConnect.us

Networking Opportunities

Training and Development

Customized Support and Mentoring

Distribution Statement A – Approved for Public Release, Distribution Unlimited 25

Pop-UpsDARPAConnect Full day events covering the breadth of your DARPA journey

• Engaging DARPA Program Managers

• Heilmeier Catechism: Understanding Effective DARPA

Communication

• Understanding DARPA Announcements and Contract Vehicles

• Reviewing and Analyzing a DARPA Opportunity

• DARPA SBIR and STTR Program

• Understanding DARPA Security Resources

• Preparing Your DARPA Proposal

• Tying It All Together: Strategies for Success

• Opportunities for Networking

Distribution Statement A – Approved for Public Release, Distribution Unlimited 26

Distribution Statement A – Approved for Public Release, Distribution Unlimited 27

CurriculumDARPAConnect

• Understanding BAAs

• SBIR/STTR

• DARPA 101

• DARPA Award Vehicles &

Solicitations

• Proposal Tips

• Heilmeier Catechism

• Engaging DARPA Program

Managers

• Becoming a PM

• DARPA Innovation Fellows

• Introduction to Security

Lessons Include:

www.DARPAConnect.us

Distribution Statement A – Approved for Public Release, Distribution Unlimited 28 https://learning.theari.us/products/darpa-award-vehicles-and-solicitations https://learning.theari.us/products/darpa-award-vehicles-and-solicitations https://learning.theari.us/products/the-heilmeier-catechism

SupportCustomized

VISIT DARPACONNECT.US

NAVIGATE THE DARPA

ENTERPRISE

ENGAGE PROGRAM

MANAGERS SPEAK THE HEILMEIER1 2 3

We offer a one-stop-shop to navigate the changes and opportunities at DARPA

Identify and engage PMs whose research interests align with your research

Frame your ideas using DARPA’s Heilmeier Catechism

For Your DARPA Journey

NOT FOR QUESTIONS SPECIFIC TO THIS BAA.

Distribution Statement A – Approved for Public Release, Distribution Unlimited 29

The Connect Corner

Webinars AMAs

Office Hours 1 on 1s

Join the LinkedIn Group:

Distribution Statement A – Approved for Public Release, Distribution Unlimited 30

For more information or to request assistance, please visit:

www.DARPAConnect.us

Thank You.

Distribution Statement A – Approved for Public Release, Distribution Unlimited 31

COOP Technical Overview

Dr. Lok Yan

Hardware is global and constantly evolving

Source: https://www.semiconductors.org/wp-content/uploads/2018/06/SIA-Beyond-Borders-Report-FINAL-June-7.pdf

Software is global and constantly evolving

Source: https://lwn.net/Articles/956765/

By lines changed

AMD 292544 24.6%

(None) 240464 20.2% Google 111034 9.3% (Unknown) 92811 7.8% Intel 87796 7.4% Red Hat 51234 4.3% Linaro 24935 2.1%

NVIDIA 23951 2.0%

Realtek 21078 1.8% Renesas Electronics 15237 1.3% MediaTek 14895 1.3%

SUSE 14581 1.2%

Huawei Technologies 13745 1.2% Qualcomm 12297 1.0% Oracle 10686 0.9%

IBM 10315 0.9%

Pengutronix 8656 0.7% Meta 7877 0.7% Amlogic 7711 0.6% Cisco 6739 0.6%

By changesets (None) 3134 18.1% Intel 1951 11.3% (Unknown) 1155 6.7% Google 1120 6.5%

AMD 1099 6.4%

Linaro 963 5.6% Red Hat 940 5.4% Pengutronix 718 4.2% Huawei Technologies 511 3.0%

SUSE 436 2.5%

Renesas Electronics 379 2.2% Oracle 365 2.1%

NVIDIA 336 1.9%

IBM 239 1.4%

Meta 225 1.3% Qualcomm 197 1.1% Arm 149 0.9% (Consultant) 143 0.8% MediaTek 117 0.7% Collabora 115 0.7%

Most active Linux Kernel 6.7 companies

Systems are global and constantly evolving

Source: https://www.mckinsey.com/business-functions/operations/our-insights/risk-resilience-and-rebalancing-in-global-value-chains

Anti-threat on opaque systems is a moving target

• Opaque: Hardware or software is opaque if some, but not all, information about its behavior is documented and known. Examples include proprietary designs, underspecified behaviors, and erroneous documentation. While not all behaviors are known, opaque hardware and software behaviors are deterministic.

Problem: Today’s cybersecurity solutions require continuous updates because hardware, software, systems, and threats continually evolve.

Question: How many hardware and software versions have you seen in the past year?

Acceptance Checks

Day 0

Continuously correct deployment

Tomorrow: proactive checks and continuous correction

Develop Threat-induced error identified and corrected

Today: reactive, separate, and single point checks

Deploy

Acceptance Checks

Day 0 Upgrade Threat found

Threat mitigationDevelop

Deploy

Software

Processor hardware

Secure Mathematically Assured Composition of Control Models

COOP shifts the paradigm from anti-threat to anti-error using stable physics

Distribution Statement A – Approved for Public Release, Distribution Unlimited 37

COOP guarantees correctness, not lack of threats or imperviousness to them

• Correct: Informally, a computation is correct if the output is as expected (both in time and value).

Strong proposals will formalize this definition.

Question: How many new hardware and software threats have you seen in the past year?

Premise: Device physics is stable. If software can be guaranteed continuously-correct using device physics, then software’s continuous-correctness guarantees are stable.

Side-channels are physical manifestations

• Side-channel: Physical manifestations of software running on hardware. Example types of side-channels include electromagnetic, power, timing, etc.

• Digital-side-channel: Manifestations of software running on hardware that are not physical.

Examples include, but are not limited to, cache-based storage and timing channels. (See side-channel definition.)

• Multi-modal-side-channel: A combination of one or more distinct side-channels plus zero or more digital-side-channels.

Premise: Using side-channels to link software execution with device physics could realize continuous correctness

An example embodiment where COOP is an interposer

COTS: Commercial off the shelf

COOP hardware and software will create a new paradigm of continuous-correctness Distribution Statement A – Approved for Public Release, Distribution Unlimited 40

Threat model for COOP interposer embodiment: Control boundaries

COOP interposer is a physical and logical reference monitor

• Can ensure data integrity of software images

(including MCS) during load

• Can mediate power delivery to processor

• Can mediate communications between processor and peripherals

• Can have additional functionality and assumptions depending on refined threat model if proposed

Proposals should refine the model and provide succinct informal logical arguments on the merits of their approach using the models

Threat model for COOP interposer embodiment: Threats

COOP is computational integrity research

Untrusted opaque software and processor

• Can violate MCS correctness (e.g., generating the wrong output)

• Cannot mount a Denial-of-Service (DoS) on the system unless the DoS is newly introduced by solution

• Cannot exceed the physical limitations of the components (e.g., assume a 1TB/s out-of-bands communications channel)

• Cannot render the component commercially unviable (e.g., randomly altering data makes component unreliable and commercially unviable)

Trusted does not mean error free

COOP solution boundaries

Within a processor package, but not on the same die.

Within a computer case (or equivalent), but not outside.

COOP solution

Processor die

Processor lid

Program identified technical challenges (from BAA)

Provable physics-based software error isolation Continuous provable error correction

Provably isolating errors in mission critical software execution not only requires differentiating correct and incorrect outputs from the MCS, but also differentiating them from other software that are executing within the opaque processors.

Errors propagate through a system exponentially with time. A single bit error could propagate to the entire system in a single second if left uncorrected.

Potential approaches include, but are not limited to, side-channel informed program synthesis, provable de-interleaving, and hardware sensor arrays.

Potential approaches include, but are not limited to, verifiable error-free checkpointing, speculative correction, and error-tracing for re-execution.

Potential approach #1: Reconfigurable hardware arrays

24 hours to capture side channels

Am pl itu de

(d Bu

V)

21 mm

Commercial chip

New GASP probe

24 mm GASP: Grid Array Side-channel Probe

Distribution Statement A – Approved for Public Release, Distribution Unlimited 45

Potential approach #2: Provable checkpointing or X I ran the software correctly

Prover VerifierClaim 1

Verify and Checkpoint

Continue

Prover-verifier formal methods

Claim 2

Restart Verify and Checkpoint

Claim 2’

Verify and CheckpointContinue

X

Distribution Statement A – Approved for Public Release, Distribution Unlimited 46

Potential approach #3: Side-channel enhancing transformations

ADD

-2

SHIFT

-2

ADD,SHIFT

-2

ADD,NOP,NOP,SHIFT

-2

1 2 3 4 5 6 7 Clock Cycle

Am pl itu de

(m V)

Source: N. Sehatbakhsh, et al. “EMSim: A Microarchitecture-Level Simulation Tool for Modeling Electromagnetic Side-Channel Signals.” In Proceedings of IEEE Symposium on High-Performance Computer Architecture, 2020.

Metrics (from BAA) H5

Metric Phase I (18 months) Phase II (18 months) Multi-threaded cores supporta In simulation In hardware

Correction latency < 100 ms < 1 ms Performance overheadb < 5x < 2x Error detection accuracy 99.9999999%c

False positive < 0.01%d

False negative < 0.05%d a Correctness techniques supports multi-threaded cores. COOP will demonstrate in simulation in Phase 1 then real multi-threaded core hardware in Phase 2.

b Compared to processor’s native performance.

c Equivalent to at least 30-bits guarantee that computation is correct; the probability of not identifying an error is 1

. This will be formally verified.

d Assessed using IV&V designed test cases.

Distribution Statement A – Approved for Public Release, Distribution Unlimited 48

• Clearly describe your threat model

• Clearly describe your control boundaries

• Clearly describe assumptions

• Clearly describe your solution

• Clearly describe new attack surfaces and mitigations (if any)

Make sure proposals are clear

Reminder: Only the information/instructions in the BAA counts (See DARPA Contracting Overview slides) www.darpa.mil

Distribution Statement A – Approved for Public Release, Distribution Unlimited 51

COOP Technical Overview: Backups

Program schedule

Supplemental options (Option 2 and Option 3)

Option 2 should include pricing for a 6-month detailed design option to bring the COOP solution to tapeout (i.e., GDSII only).

Option 3 should include pricing for a 7-month COOP fabrication and packaging effort.

Contractor designed, but government provided fabrication and packaging

Contractor designed and contractor provided fabrication and packaging

Proposals must clearly describe the oracle used

• Oracle: An abstract entity used in formal methods to represent the axiomatic source of correct answers. Formal specifications and golden models are examples of oracles. Proposals must clearly describe the oracle used, why the oracle is realistic, and mitigations for if and when the oracle is proven wrong in the future.

Error detection accuracy of 99.9999999% will be formally verified

• Equivalent to at least 30-bits guarantee that computation is correct

• Probability of not identifying an error when there is one is 1

In the minimum:

• Consider how the problem is formalized

• Consider what assumptions are needed

• Consider how the guarantees can be proven

• Consider who can verify proofs

MTO Overview

Dr. Whitney Mason www.darpa.mil

Distribution Statement A – Approved for Public Release, Distribution Unlimited 58