CSO QA Response.pdf
PDF 101 KB Posted
- Attached to
- Call 001: Data Models and Flows Under CSO FA9302-21-S-C001 Federal contract opportunity
- Solicitation number
- FA930021SC001Call001FY25
About this file
This is a Q&A document providing detailed responses to industry questions about a Department of Air Force contract opportunity for RF data analytics and visualization capabilities. The work involves developing automated AI/ML analytical applications to process In-Phase and Quadrature (I&Q) data sets ranging from 10GB to 1TB, with the systems operating in Top Secret SCI & SAP environments.
Key technical requirements include: processing recorded (not live) signals within 24 hours, handling both defensive and offensive analytics with an initial focus on defensive capabilities, supporting multiple I&Q data formats (complex32, complex64, float32, float64), and operating in both on-premises Windows servers and AWS-based VMs in the Top Secret Region of FENCES. The work will be performed primarily at Edwards AFB (CA), Nellis AFB (NV), Eglin AFB (FL), China Lake NAS (CA), Wright-Patterson AFB (OH), and National Capital Region locations. This is a new requirement without incumbent contractors. The Multi-Domain Test Force will be the primary point of contact for requirements and performance evaluation. The government will provide test equipment including spectrum analyzers and antennas. Contractors must develop applications that will be government-owned, with one dataset expected per month for analysis.
View the file
Other files for this federal contract opportunity
| File | Type | Posted |
|---|---|---|
| 25-MDTF_Data Analytics Support_CSO_Call 001.pdf |
On GovTribe
Work with this file on GovTribe
- Download the original file
- Contacts named in this file
- Similar government files
- Ask GovTribe AI about this file
Text version
CSO Q&A
Data Models and Flows:
Where will the data model(s) and flow(s) be hosted? Does AFTC plan to host the models and flow locally, on-base in a data center, or is AFTC planning on using a cloud service provider (CSP)? That is part of the expected proposal. Currently data flows to a workload in the Top- Secret region of FENCES.
What Impact Level(s) are the model(s) and flow(s)? Please list all Impact Levels needed. The models will be used on Top Secret SCI & SAP datasets.
Does AFTC require Hybrid cloud solutions? Hybrid cloud solutions are defined as a combination of on-base infrastructure networked to cloud-based services. That is part of the proposal, but that would be a reasonable solution.
Radio Frequency Analytics and Visualization:
Describe what type of Artificial Intelligence / Machine Learning (AI/ML) services a/o applications MDTF requires? Are these GOTS or COTS? The applications development is the core of the contract – contractor developed, but government owned.
If GOTS a/o COTS AI/ML is planned, please list the GOTS/COTS applications and the technical specifications required to host in the cloud. This is the proposal we are expecting to get from vendors.
Where will the AI/ML be hosted? Does AFTC plan to host the models and flow locally, on-base in a data center, or is AFTC planning on using a cloud service provider (CSP)? That is part of the expected proposal. Currently data flows to a workload in the Top-Secret region of FENCES.
What Impact Level(s) are the RF and Visualizations? Please list all Impact Levels needed. Top Secret SCI & SAP.
If working with a third-party contractor is something you’d consider for this requirement? Yes.
Could you please clarify if the place of performance is at Edwards AFB, or if it is in Las Vegas, NV? Will the services be performed at Government or contractor facilities? Primarily at or near the following locations: Edwards AFB California; Nellis AFB Nevada; Eglin AFB Florida; China Lake NAS California; Wright-Patterson AFB Ohio; National Capital Region (various locations)
This this a new requirement, or is it a continuation of prior work? If a continuation of prior work, can the Government please provide information on the prior contracts? This is a new requirement
Is this effort focused more on development of an automated I/Q processing system, or on the engineering services to run a commercial product, or both? More on development of an automated I/Q processing system.
Is the I/Q data a single format, or is it in multiple formats? Is the IQ data in an open standard or openly available data format? Multiple formats: All data have an In-phase and Quadrature component, but the encodings vary (e.g. complex32, complex64, two float32, two float64, etc) and the file formats vary.
Has the Government identified any external systems that this IQ processing system must interact with? Systems are not limited to, but include Windows Server based on-premises systems and various AWS-based virtual machines in the Top Secret Region of FENCES.
Can the Government please provide information on tools currently in use? No, this is a new requirement.
Is the Government equally interested in processing comms and radar signals, or is one more important than the other? Equally interested.
Is the expectation for the AI/ML models to function in the cloud, on prem, and/or DDIL?
Cloud (Top Secret Region of FENCES) and on-prem systems.
Is the expectation of use of these models primarily defensive (IE for the purpose of seeing noisiness of currently in use hardware) and/or offensive (IE for determining data about threats)? Expectation is to be able to handle both but start with defensive.
Is the expectation to have the RF sensors be human carriable? The RF sensors vary from handheld to vehicle installed.
Will the solution need to analyze live signals or just recorded ones? Just recorded ones.
What is the speed requirement for analyzing signals? Within 24 hours.
If an acceptable machine learning model cannot be found, can you provide training data to create a new machine learning model? Yes.
Will the analyzed signals be analog and/or digital? Will they be voice or data? Will they be noise or RFI? Digital, encoded data.
Will signals need to be analyzed as a set, like in radio direction finding? Each I&Q file will need to be analyzed, regardless of content.
Systems & Infrastructure
Existing Systems & Data Sources
What existing data management and analytics systems are currently in place for RF data processing? This is a new requirement, and that is part of the expected proposal.
Are there any legacy systems that need to be integrated or phased out as part of this effort?
No.
What types of storage and computational infrastructure (e.g., HPC clusters, cloud services, on-prem servers) are available for handling large datasets? On-prem servers and various AWS-based VMs in the Top Secret Region of FENCES.
System Architecture & Interoperability
What operating environments will this system need to support (e.g., on-premise, classified networks, cloud-based DoD environments)? On-prem servers and various AWS-based VMs in the Top Secret Region of FENCES.
Are there existing Air Force or DoD enterprise architectures or frameworks that the solution must align with? On-prem Windows Servers based systems and various AWS-based VMs in the Top Secret Region of FENCES.
Will the system need to support real-time data ingestion and processing, or will it primarily handle batch processing? Post-collection batch processing.
Data Sharing & Collaboration
Will data from different sources (e.g., radars, satellites, communication systems) need to be combined or cross-analyzed? Yes.
How will data be shared between different agencies and branches (e.g., USAF, Navy, Army, Space Force)? Via existing means. The expectation is for the performers to provide the technical solutions.
Are there any restrictions on data transmission or storage locations due to security classification levels? Yes. Expect to work in a TS/SAP and/or TS/SCI environment for analysis.
Stakeholders & Decision-Making Process
Previous & Current Stakeholders
Have there been prior vendors or contractors working on similar RF data analytics solutions? If so, what challenges were encountered? This is a new requirement.
Who are the primary government stakeholders involved in defining requirements and evaluating performance? The Multi-Domain Test Force is the focal point for requirements and evaluating performance.
Are there any third-party vendors or technology providers that the Air Force prefers to work with for analytics and visualization? No.
User Roles & Expectations
Who will be the primary end users of the system (e.g., analysts, field operators, mission planners)? Analysts.
What level of technical expertise do the users have in data analytics and visualization? A high level.
What level of automation is expected versus manual analytical review? That is part of the expected proposals, but more automation is desired.
Feedback & Iterative Development
Will there be opportunities for direct feedback from stakeholders during development? Yes.
What level of flexibility exists for iterative improvements and refinements based on testing results? That is expected to be part of the proposal, but the ability to work in the TS/SAP and/or TS/SCI environment would increase the value of an iterative process.
Are there existing frameworks or processes for Agile or DevSecOps development within the Air Force? A Git lab with CI/CD pipelines is available at the Unclassified level, and Git repo’s are available on the high side. A Data Transfer as a Service function is available to go from low to high (but not high to low).
Cross-Branch & Interagency Coordination
Involvement of Other Military Branches & Agencies
Will this effort require coordination with other branches such as the Navy, Army, Space Force, or NSA? MDTF is the primary Point of Contact.
Are there existing data-sharing agreements between agencies that impact access to relevant datasets? Any data-sharing issues are the government’s responsibility.
Standardization & Compatibility
Are there standardized formats, frameworks, or interoperability requirements mandated by the DoD for RF analytics systems? No.
What levels of data classification and dissemination controls need to be followed across different military branches? Data Classification and dissemination is the government’s responsibility.
Are there cross-domain security solutions required for transferring data between classified and unclassified environments? A Data Transfer as a Service function is available to go from low to high (but not high to low) once inside the FENCES environment.
Logistics, Procurement, and Operational Considerations
Testing & Deployment Logistics
Will testing and deployment occur at specific Air Force or DoD facilities (e.g., Edwards AFB, Eglin AFB, Nellis AFB)? Yes, and in the Top Secret Region of FENCES.
What are the physical and logistical requirements for deploying and maintaining this system in an operational setting? That is part of the expected proposal, but on-prem and cloud based systems (FENCES) are available.
Are there existing test ranges or environments where the system will be validated?
Validation will occur in the Top Secret Region of FENCES and on-prem systems.
Training & Knowledge Transfer
Will the contractor be responsible for training military personnel and analysts on the system? If the delivered capability is not intuitive, then yes.
What documentation, user guides, or training materials will be required? If the delivered capability is not intuitive, then sufficient information to make the capability useful.
Is there an expectation for ongoing training and support post-deployment? No.
Supply Chain & Procurement
Are there any restrictions on sourcing hardware or software due to DoD supply chain risk management (SCRM) policies? Sources must be compliant with DoD policies and regulations.
Are there any specific vendors or technologies that must be used due to existing government contracts? No.
What is the procurement timeline for any required third-party tools, hardware, or software? Timelines are part of the expected proposal.
Field Deployment & Scalability
Will the system be required to operate in austere, disconnected, or edge environments (e.g., deployed units, aircraft, ships)? No.
Are there scalability requirements for handling larger datasets or higher operational tempo in future conflicts? No, data sets are test vs operationally focused.
Will there be any mobile or lightweight versions of the analytics tools required? No.
Funding Availability & Allocation
Are there any preliminary budget estimates or funding ceilings that we should be aware of when structuring our Rough Order of Magnitude (ROM) estimate? The ROM must be an estimate of the cost to design build and delivery the new technology to the government. There will be no budget estimates or funding ceilings provided as a part of this call.
Will funding be allocated in phases, and if so, what are the expected milestones for funding disbursement? Contract CLIN structures and obligation of funding will be a part of the second phase of the CSO, once a white paper has been determined to be in the best interest of the government and we move forward.
PARAGRAPH REFERENCE - 2.5.1
Will the RF information be encoded in stored bus data format? Is it in need of transformation and/or decoding for usability to engineering units? Yes and yes.
PARAGRAPH REFERENCE - 2.5.1
Can the govt provide examples of parametric signals being recorded? Yes, at the appropriate classification levels.
PARAGRAPH REFERENCE - 2.5.1 & 2.5.2
What are the frequencies of signals being recorded? Is there a need for real-time processing and live-feed analytics? The analytical tools are expected to be applied to I&Q recordings with the center frequency captured in meta data, so the analytical tools should not be limited to a specific frequency range. No requirement for real-time nor live-feed analytics.
PARAGRAPH REFERENCE - 2.5.2
What types of decisions can and/or need to be made using the signal information? It is to analyze test data for procurement decisions.
PARAGRAPH REFERENCE - 2.5.1
Where will the data model(s) and flow(s) be hosted? Does AFTC plan to host the models and flow locally, on-base in a data center, or is AFTC planning on using a SAP/SAR cloud service provider (CSP)? If a CSP, does the government have preference for a specific environment?
Yes, there are on-prem Windows-based systems and various AWS-based VMs in the Top Secret Region of FENCES.
PARAGRAPH REFERENCE - 2.5.1
What is the frequency of the dataset (10Gb to 1Tb) expected? Daily, Weekly, etc. Looking to understand the total data storage needs and how much of the data is hot, cold, or archive Expect one dataset per month.
PARAGRAPH REFERENCE - 5.6.1
Will the government please clarify the Government Furnished Property for this requirement?
This will assist with pricing with "transportation, installation, or rehabilitation costs". Primarily test equipment, to include but not limited to, spectrum analyzers, antennas, and equipment to capture data for post-test analysis.
File details come from the government source that posted it. Updated .