HR001118S0004-Amendment-02.pdf

Amendment 02 to DARPA BAA HR001118S0004, Geospatial Cloud Analytics (GCA)

The purpose of Amendment 02 to DARPA BAA HR0001118S0004, Geospatial Cloud Analytics (GCA) is to make the following change:

A. Incorporates Questions and Answers (Q&A)

1. Is text-mining or other non-traditional GEOINT sources within the scope for GCA?

a. Yes, as seen in Section 1.1 of the BAA, non-traditional data sources can be used in the GCA program.

2. The cloud offering must be available on what DoD network levels?

a. Per Section 1.3 of the BAA, GCA is an unclassified program.

3. Can you break down the importance of the evaluation criteria in terms of percentages?

a. Per Section 1.5 of the BAA, The BAA lists the evaluation criteria in descending order of importance but there are no percentages given to the criteria.

4. Does GCA have access to commercial data that other Government agencies have purchased?

a. Any geospatial data sources freely available to Government contractors or performers may be used in the program.

5. Is there a downselect from Phase 1 to Phase 2?

a. There is no downselect from Phase 1 to Phase 2. However, per Section 1.2 of the BAA, all proposals should have a Phase 1 base with a Phase 2 option and it is the Government’s decision whether or not to exercise any proposal option.

6. What type of security should TA-1 have?

a. It is up to the proposers to decide what security their platform will have.

However, proposers should refer to Section 6.2.2. for additional guidance on security.

7. What’s your definition of “near real-time” accessibility?

a. Near real time accessibility is dependent on the requirements of the problem area to be addressed.

8. What cloud provider should be chosen?

a. It is up to the proposer to choose and justify an appropriate cloud provider given the details provided in Section 1.2.4 of the BAA.

9. Who is the transition partner for GCA?

a. On a macro level, the program intends to work with National Geospatial

Agency. However, a goal of the program is to help underserved lower echelons e.g. (COCOMS and below). Performers are also welcome to bring in potential transition partners.

10. Who were the SBIR (16.2 Software/Analytics Exploiting Commercial Satellite Imagery) performers you referenced and what were their problem areas?

a. There were 7 performers on the Commercial Satellite SBIR.

i. Applied Research – A Novel Border Monitoring System Using Medium and High Resolution Commercial Satellite Images

ii. Descartes Labs – Satellite Imagery Analysis for Automated Global Food Security Forecasting

iii. DZYNE – Cloud Based Analytics of Satellite Imagery

iv. Heureka – Leap Frogging the State of the Art in Satellite Exploitation

v. LongShortWay – Wide Area Sensor Fusion

vi. OmniEarth – Precision Soil Moisture Mapping Based on Fusion of SAR with Unconventional Data Sources

vii. Vision Systems – Rapid, Reliable, and Global Analytics using Commercial Satellite Imagery

11. Are you open to a demo prior to BAA submission?

a. While an in-person demo is not available due to the BAA being released, proposers are welcome to include a description of their system within the page limits allotted, as well as include a link to an online demo in their proposal.

12. Can derived products be included on the TA-1 platform?

a. It should be noted that the primary focus of GCA is on analytical services that provide products for the military. However, it is up to TA-2 performers whether they would like to share their results and up to TA-1 performers on whether they want to include those results in their platform.

13. Can classified data sources be proposed?

a. Since lower echelons often do not have access to classified geospatial data, GCA is an unclassified program (Section 1.3 of the BAA) and therefore we do not anticipate using classified sources.

14. How should TA-2 performers cost out cloud compute resources?

a. Performers should use costing tools that are publically available through most cloud providers. A ‘most likely’ cloud provider or an average of the most likely providers can be used.

15. Does TA-2 need to demonstrate their algorithm on a global scale?

a. No. Per Section 1.1. of the BAA, the TA-2 software will be proven out at a lower level and global scalability will be proven out at a mathematical level.

16. Can multiple open topics be proposed?

a. Yes as long as the page limits in Section 4.3.2 are adhered to.

17. Do TA-2 performers pay for data licensing, access, etc?

a. No, TA-1 pays for data licenses and coordinates licensing for users on their platform.

18. What is the level of data curation and processing expected from TA-1 providers?

a. It is expected that during Phase 1, TA-1 performers will work with TA-2 performers to negotiate at what level the data should be curated and processed.

19. What level of maturity is allowed?

a. Due to the quick nature of the program, a reasonably mature system is expected at the outset.

20. Do all of the data sources have to be global?

a. No, any geospatial data is allowed on the platform but proposers should note that there may be a risk of not being able to scale globally if the data source is only available for a certain area of the world.

21. Are non-geospatial data sources allowed?

a. Yes, as seen in Section 1.1 of the BAA.

22. Can a TA-2 solution span multiple problem areas?

a. Yes. However, all metrics for TA-2 (Section 1.4 of the BAA) must be met.

23. Are TA-2 performers expected to keep up with the TA-1 ingest rate?

a. No, TA-2 performers only have to adhere to the TA-2 metrics.

24. Are foreign data providers and foreign partners allowed?

a. Yes, we anticipate foreign data sources being used in the program. Note that foreign data purchases are distinct from foreign partnerships, which are discussed in Section 3.1.2.

25. If a small business is bidding as a prime proposer are they still required to submit a plan with their proposal?

a. All proposers are required to adhere to the instructions of the BAA.

26. Section 5.1.4 refers to proposer’s capabilities and/or related experience, where in the technical volume should this information be included?

a. It can be included anywhere in the technical volume but the Detailed Proposal

Information section is preferred.

27. Would the Government consider excluding the proposed statement of work from the

Section II Detailed Proposal Information page limit?

a. The proposed Statement of Work is included in the detailed proposal information page count (Section 4.3.2). The Government will not be amending the BAA to make any changes.

28. Should TA-1 provide reference data during Phase 1 that can be used by TA-2 performers to assess performance?

a. It is up to the TA-1 performers whether they want to provide example data during Phase 1, noting, however, that it is beneficial for TA-2 performers to get access to the platform during Phase 1 to validate the API.

29. When or how would this project become subject to ITAR restrictions?

a. It is the performer’s responsibility to comply with ITAR requirements.

30. Can an organization that is a subcontractor on a TA-1 proposal submit a proposal for

TA-2?

a. Yes.

31. What is the role of a TA-2 performer during Phase 1?

a. As seen in Section 1.2.1 of the BAA, TA-2 performers should provide feedback and suggestions for the APIs that TA-1 performers are creating.

They may also give feedback and suggestions on the data sets provided, levels of curation/processing, or any additional features that may make the platform more useful. Additionally, some work may begin on algorithm development and scaling.

32. Does DARPA intend to post the Proposer's Day presentation on a website?

a. The Proposer’s Day presentation is posted on the Federal Business

Opportunities website (www.fbo.gov).

33. 1.4. PROGRAM METRICS, Table 9 lists SAR as a minimum data source. Will the government accept foreign SAR data since no commercial US companies do this?

a. Yes.

34. 1.4. PROGRAM METRICS, Table 9 states: "Ability for platform to sustain 10 or more data sources (3 commercial)..." Can all 10 sources be commercial? Or does the Government plan to supply 7 of the data sources? Will having a variety of imagery and non-imagery sources be in the favor of TA-1 proposers?

a. The BAA does not preclude the use of more than three commercial sources.

b. DARPA does not intend to supply any data, but the US Government and other organizations do freely offer certain data to organizations that request it.

c. TA-1 proposals should meet the metrics listed for TA-1. It should be noted that a varied data set may help attract more TA-2 performers. Examples of RF data sources also include VMS, AIS, and ADS-B."

35. 1.4. PROGRAM METRICS, Table 9, Data ingest rate states "Make new data available to TA-2 performers at an average of > 200 TB/day." Is this requirement for Phase 1 and Phase 2?

a. The TA-1 platform should be able to sustain data ingest and preprocessing at those rates. At a minimum, the performer should show, through analysis and extrapolation, the ability to scale to that rate.

36. 4.3.2. Proposal Format, 4.3.2.1 Volume I, Technical and Management Proposal, section II F states that "A three slide summary of the proposal in PowerPoint" is to be included. Can we assume that these slides do not count towards the overall page count and will be submitted as a separate PowerPoint file that is included in the overall zip file?

a. Section 4.3.2.1 answers this in detail.

37. 6.2.2. "Protecting Controlled Unclassified Information in Nonfederal Information

Systems and Organizations" (Section 6.6.2) If all of the data and algorithms are commercial products, will any of the data be considered Controlled Unclassified Information (CUI)? If any data will be CUI, will it be marked and identified by the government in advance?

a. It is the responsibility of the performers to identify any CUI or otherwise restricted data in their proposals.

38. If a proposer has worked with Government agencies such as SPAWAR or various COCOMs on relevant technologies, can they be listed as transition partners?

a. Yes. All performers are welcome and encouraged to bring Government transition partners to the table.

39. Does the government intend to request or provide preferred data sources or is the intent for each development team to identify and acquire 100% of the data sources? If vendor is responsible for identifying all data used, would the team be able to ask front-line analysts outside of the BAA teams for input on actual use cases?

a. It is up to the TA-1 performers to choose the data sources that they will provide on the platform and TA-2 providers to request data sources. During Phase 1 additional discussions may be had between TA-1 and TA-2 to negotiate the final data sets on particular platforms. Asking transition partners about preferred data sources is allowed and encouraged.

40. What is the requirement for system availability during Phase 2 for TA-1 vendors?

Must the system be available all the time or is it required that the TA-1 team coordinates with TA-2 teams to ensure that the TA-1 team is not inhibiting TA-2 teams' development?

a. It is up to the TA-1 performer to ensure that TA-2 performers have adequate access to their platform for testing and development.

41. What is the envisioned implementation or integration of these systems? Will TA-1 vendors develop the solution in a commercial cloud environment? Does the commercial cloud provider need to be DISA approved? Will the system need to be built in a government cloud on a government network?

a. Refer to the BAA for specific details of the program, including security requirements.

42. While this system will be focused on unclassified government and commercial data, does the government anticipate that log files, access patterns, user account information, etc., taken in aggregate, will require a higher level of protection? What is the anticipated DISA impact level of the system including aggregate metadata?

a. It is anticipated that the GCA program will be unclassified, and that systems will provide protections similar to those used in commercial systems.

43. Will the Government allow exclusive teaming of commercial data providers for TA-1 and TA-2?

a. The conditions governing teaming requirements are in the BAA. It should be noted that TA-1 performers are expected to collaborate to make it easier for analytics programmers to port their work and algorithms among the platforms.

Exclusive teaming between data providers and TA-1 performers is strongly discouraged as it may limit access byTA-2 performers. Additionally, TA-1 performers will be evaluated on several factors including the ability to support the selected TA-2 performers.

44. What is the "state of the art" the TA-2 app will be compared with as referenced in Tables 12 & 13 Metrics for Phase 2 competitions?

a. The state of the art varies based on problem area. The proposer may choose to identify these in their proposal.

45. How does the Government define the size used in scaling assessment? (Referencing Table 11 Implementation Scalability) Is this number of areas covered or dimension of areas covered? Or is it up to app to decide/define in approach?

a. The objective is good weak scaling. The reference to a problem of size X refers to the size of the input data.

46. Does the government prefer an Other Transaction type of contract and if so, is DARPA interested in waiving requirements for CAS compliant companies in order to qualify?

a. BAA Part 1 states that DARPA may award procurement contracts or other transactions. However, there is no preference stated.

47. Are TA-1 platforms that require internet connectivity to run properly not considered viable solutions for the GCA program?

a. The GCA program assumes that Internet connectivity will be required to run in most scenarios.

48. For TA-2: Are we designing for the current commercial imaging constellation or a future one? If future, how much more coverage is appropriate to assume?

a. The TA-2 performers should design to either currently available data sources or data sources that will be available within the anticipated period of performance. It should be noted that implementation scalability is a TA-2 metric, and therefore the ability to scale to much greater coverage in the future is desirable.

49. For TA-2: Is the re-tasking of satellite assets in-scope or should we focus on exploiting existing data?

a. Re-tasking of satellites is not in scope.

50. For TA-2: Can two parts of the same company submit separate proposals to TA-2?

a. Yes.

B. All other original DARPA BAA HR001118S0004 data remains unchanged.