JA_VISN_6_BioMed- Final_Redacted.pdf

CONTROL NUMBER VA-26-00083446

JUSTIFICATION FOR AN EXCEPTION TO FAIR OPPORTUNITY

1. Contracting Activity: Department of Veterans Affairs Office of Procurement, Acquisition, and Logistics Technology Acquisition Center 23 Christopher Way Eatontown, NJ 07724

2. Description of Action: This proposed action is for a firm-fixed-price delivery order to be issued under the National Aeronautics and Space Administration (NASA) Solutions for Enterprise-Wide Procurement (SEWP) V Governmentwide Acquisition Contract (GWAC) for brand name NVIDIA Graphics Processing Units (GPU) and software, Cisco Systems servers and management software, and F5 Networks Application Delivery Controllers with annually renewable premium Original Equipment Manufacturer (OEM) service, with Optional Tasks for brand name Cisco Fabric Interconnects, additional Cisco/NVIDIA GPU compute nodes, NetApp storage expansions, and NetApp support renewal.

3. Description of the Supplies or Services: The Department of Veterans Affairs (VA), Office of Information and Technology (OIT), Infrastructure Operations (IO) has a requirement to provide dedicated GPU-compute capability, refresh end-of-life Application Delivery Controllers, and expand storage within the existing converged Biomed data center architecture at seven Veterans Integrated Service Network (VISN) 6 facilities (Asheville, Durham, Fayetteville, Hampton, Richmond, Salem, and Salisbury).

The compute, storage, and application delivery environment at these sites is Government-owned and hosts the TeraRecon clinical diagnostic imaging workload. To ensure native integration with the existing architecture and clinical applications, the Contractor shall provide the following base requirements:

Cisco Unified Computing System (UCS) Hardware & Software: Fourteen (14) Cisco UCS C220 M8 Rack Servers, two (2) per site, forming the minimum viable GPU compute cluster at each facility which includes Cisco Intersight Software as a Service license. Cisco hardware warranty is covered under a separate, existing VA Enterprise Software Agreement (ESA).

F5 Networks: Fourteen (14) F5 BIG-IP LTM R2800 Load Balancers, two (2) per site, each with OEM Warranty and Maintenance Support and an F5 premium maintenance and support service providing service-level-backed 24x7 support, expedited hardware replacement, and BIG-IP software support.

NVIDIA Hardware & Software: Fourteen (14) NVIDIA L4 24 gigabyte GPUs, one (1) per server, paired with NVIDIA Quadro Virtual Data Center Workstation (vDWS) concurrent user licenses.

Additionally, VA has included optional tasks for the following:

- Up to fourteen (14) 6th GEN UCS Fabric Interconnects (UCS-FI-6652-U) with perpetual software licenses, UCS Manager v6.0, and required accessories.

Hardware warranty is covered under the existing VA ESA.

- Up to fourteen (14) 50 terabyte (TB) Non-Volatile Memory Express (NVMe) drive packs. Each pack must include all necessary components, including NVMe drives, enclosures/shelves, cabling, and software licensing required to integrate the capacity into the existing NetApp AFF C250 appliances at each location.

Warranties must be co-termed with the existing C250 appliances. If Optional Task 2 is exercised, NetApp storage NVMe drive pack expansions shall include OEM warranty and maintenance support and have their warranty and maintenance support co-termed with the existing NetApp AFF C250 appliances currently located at the VISN 6 facilities.

- Up to twenty-five (25) additional Cisco UCS C220 M8 servers with NVIDIA L4 GPUs and associated vDWS and Intersight subscriptions, identical to the base configuration, exercisable per site up to stated per-site maximums. The associated NVIDIA vDWS 1CCU and Cisco Intersight SaaS Essentials subscriptions may be renewed, if needed, for Option Periods 2 & 3.

- OEM warranty and maintenance support renewal for the seven (7) existing Government-owned NetApp AFF C250 appliances beginning at the expiration of existing pre-paid support (August 31, 2027). For the additional quantities exercised in optional task two, coverage periods shall co-term with the existing support, pro-rated up to twelve (12) month increments, exercisable per appliance. Maintenance may be renewed if needed for Option Periods 2 & 3.

The Contractor shall perform site readiness activities, and deliver, install, and configure hardware within 90 days of award.

The base period of performance shall be 12-Months from the date of award with all hardware delivered, installed and configured within 90 days. Also, Two, 12-month option periods for continues Licensing and support and four optional tasks for Cisco Fabric Interconnects, NetApp Storage Expansion packs, Additional GPU nodes and NetApp Support renewal. The total estimated value of the proposed action is inclusive of all option periods and optional tasks.

4. Statutory Authority:

Pursuant to Revolutionary Federal Acquisition Regulation Overhaul (RFO) 16.507- 7(a), the following brand name requirement is being justified and approved as an exception to fair opportunity under the statutory authority of Section 41 U.S.C.

4106(c)(2) as implemented by RFO 16.507-6(b)(2), entitled “Only one awardee is capable of providing the supplies or services required at the level of quality required because the supplies or services ordered are unique or highly specialized.”

5. Rationale Supporting Use of Authority Cited Above:

The proposed action is to procure brand name NVIDIA GPUs and software, Cisco servers and Fabric Interconnects, F5 Networks Application Delivery Controllers with premium OEM service, and NetApp storage expansions and support renewal. This action extends the current VISN 6 Biomed data center architecture was designed, procured, and operated by OIT as a single, unified converged solution. Each brand name product is essential to the Government’s requirements, and market research indicates other companies’ similar products, or products lacking these features, do not meet, or cannot be modified to meet, the agency’s needs. These components must seamlessly integrate into the existing VISN 6 Biomed virtualization and application delivery architecture and its established management plans. Introducing competing OEM hardware or software into this environment would break the current architecture, introducing critical configuration incompatibilities, massive technical debt, and an unfeasible engineering overhaul. The justification for each individual brand is provided below, followed by the justification for procuring the items as a single, unified solution.

Cisco Servers and Fabric Interconnects: The VA utilizes Cisco Intersight and Cisco UCS Fabric Interconnects to deploy, manage, and scale server compute clusters centrally. Alternative server hardware (such as Dell or HPE) cannot be integrated into the existing Cisco UCS management domains or the deployment mode for UCS Fabric Interconnect connected servers in UCSM mode. Because Cisco’s proprietary technology handles firmware deployment and service profile management across VA’s infrastructure, introducing a third-party server would break uniform policy enforcement across the enterprise. Additionally, Cisco hardware warranty is already covered under a pre-existing VA Enterprise Software Agreement (ESA); procuring any other brand of server would require the Government to purchase duplicate warranty and support coverage it already owns.

Virtualization Environment Constraint (no hypervisor procured): The compute delivered under this action must be certified for, and integrate natively into, the Government’s existing virtualization environment and its management plan at these sites, without disruption to the running clinical workloads. This constraint narrows the field of acceptable server, GPU, and driver combinations to those certified on the compatibility matrices of the installed environment and the clinical application Internal Standard Variation (ISV), and it is a controlling factor in the brand determinations below.

NVIDIA GPUs and vDWS Software: The TeraRecon application vendor certifies its virtualized diagnostic imaging delivery on NVIDIA vGPU technology; competing accelerators are not certified by the ISV for this clinical workload and are not licensable within the existing NVIDIA Quadro/vDWS concurrent license pools. In addition, Government engineering validated the delivered configuration against OEM published configuration rules and Government-measured site environmental data collected from all thirty-three (33) existing hosts across the seven sites in July 2026;

the NVIDIA L4, as a 70-watt, passively cooled, single-slot accelerator, is the GPU class the OEM supports in the required 1U server configuration for the sites’ measured environmental envelopes.

F5 Networks Application Delivery Controllers: VA heavily leverages F5 BIG-IP appliances for Local Traffic Management (LTM), managed through F5’s proprietary BIG-IQ central management and uniquely customized using F5’s proprietary “iRules” scripting language. An alternative brand would not recognize these iRules, disabling essential application routing for health care portals.

NetApp Storage Expansions (Optional Task): The VISN 6 facilities currently operate existing NetApp AFF C250 storage appliances. To add 50TB capacity expansions to these arrays, the NVMe drive packs and shelves must be manufactured by NetApp.

Third-party disk drives or enclosures will not be recognized by the NetApp ONTAP operating system, cannot be managed by the existing NetApp controllers, and would void the existing hardware warranty of the C250 chassis. The warranty must be co-termed directly with NetApp to ensure continuous, unified support. Likewise, OEM support renewal for the existing C250 appliances (Optional Task 4) can only be provided by NetApp: third-party maintenance cannot deliver ONTAP software updates, security patches, or OEM parts entitlement, and would fragment support accountability for arrays hosting clinical imaging data.

Requirement for a Single, Unified Procurement: It is critical that these items are procured together from a single authorized reseller. These products are not standalone, disparate Information Technology (IT) commodities; they are tightly coupled components of a unified VISN 6 Biomed Virtualization, Server, Storage Expansion, and Application Delivery Refresh. The entire solution is dependent on each individual part, and the items cannot be split logically into separate procurements for the following technical and logistical reasons:

Interoperability and Bill of Materials (BOM) Validation: This hardware and software form a single, integrated virtualization stack: the Government’s existing hypervisor environment running on the Cisco compute layer, backed by NetApp storage, accelerated by NVIDIA GPUs, and load-balanced by F5. By procuring this through a single Value-Added Reseller (VAR), the reseller is responsible for validating the entire BOM prior to order execution. The reseller ensures that the specific Cisco server chassis has the exact power and riser configurations needed to support the NVIDIA GPUs, and that the transceivers and licensing match up across the Cisco fabric and NetApp storage. If the procurement is split, the VA loses that unified engineering validation and absorbs all integration risks.

Supply Chain and Deployment Synchronization: The Cisco servers cannot be configured without integration into the existing licensing and management planes, and the GPU cluster cannot serve the clinical workload without the load balancing and storage layers. Breaking this into separate procurements places the deployment at the mercy of disjointed delivery timelines. If servers arrive weeks before the remaining components, the hardware remains non-operational while consuming its warranty period. A single reseller coordinates the supply chain to ensure synchronized delivery so the stack can be built immediately, which is critical to fielding this clinical imaging capability within the 90-day installation schedule.

Warranty and Support Alignment: Procuring as a single solution ensures that the hardware support, software maintenance, and licensing terms all initiate simultaneously. This is critical for maintaining a standard enterprise hardware lifecycle without administrative gaps or lapses in support coverage.

Single Point of Accountability: Disjointed procurements that result in out-of-sync arrivals or incompatible firmware/licensing place VA engineering teams in the middle of multiple vendors deflecting responsibility. A unified procurement ensures a single point of accountability for the delivery and baseline interoperability of the solution.

Replacing any part of this solution would require a complete replacement of the current infrastructure for virtualization, compute, storage, and application delivery systems across all seven VISN 6 locations. Such a replacement would take approximately two years and an estimated to rebuild and develop.

These estimates are based on previous acquisition history. These duplicated costs would not be recovered through competition.

While the items must be purchased as a single package, market research (detailed in Section 8) confirms that there are multiple authorized Service-Disabled Veteran- Owned Small Business (SDVOSB) resellers on the NASA SEWP V GWAC capable of fulfilling this entire multi-brand BOM.

6. Efforts to Obtain Competition:

Market research was conducted, details of which are in Section 8 of this justification.

This effort did not yield any additional brands that could meet the Government’s interoperability requirements. However, the undersigned VA technical representative determined that competition is available among authorized VAR’s for the brand name hardware, including the associated software maintenance and support. In accordance with FAR 16.507-7(c), this justification will be provided with the solicitation to all appropriate NASA SEWP V GWAC contract holders. Furthermore, in accordance with FAR 5.301, and 16.507-6(f), the award notice for this action will be synopsized on the Contract Opportunities website and this justification will be made publicly available within 14 days of award.

7. Actions to Increase Competition:

The Government will continue to monitor the IT hardware and virtualization marketplace to ascertain whether future interoperability standards evolve to a point where cross-brand integrations (such as cross-platform hypervisor management or open-standard appliance replication) become natively viable without requiring parallel infrastructure. Market research, including RFIs, will be conducted before any follow-on acquisition to determine whether the requirement can be competed at that time.

8. Market Research:

Market research was conducted by VA technical experts via web-based research and historical procurement analysis from February through May 2026.

VA subject matter experts (SME) reviewed alternative products including Microsoft Hyper-V, Red Hat Virtualization/OpenShift Virtualization, and Nutanix AHV hypervisors; AMD GPUs; HPE and Dell servers; Dell EMC storage; and Citrix ADCs.

The VA SME’s determined that alternative hypervisors cannot perform live migrations into the VA’s existing virtualization clusters or be managed by the existing management plane. Alternative GPUs do not integrate into the existing NVIDIA Quadro/vDWS concurrent license pools required by the Terarecon virtual machines.

Alternative servers cannot be managed by Cisco Intersight or Cisco Fabric Interconnects. Alternative storage drives cannot be inserted into the existing NetApp AFF C250 arrays. Alternative load balancers cannot execute F5 iRules. Based on these market research efforts, the Government’s technical experts determined that only the brand name products identified herein can meet all of VA’s interoperability and compatibility requirements.

The Contract Specialist (CS) conducted additional market research by conducting an RFI on the NASA SEWP GWAC from June 11, 2026, to June 25, 2026. VA received three responses from NASA SEWP GWAC SDVOSB resellers, confirming that the full solution was available through multiple VAR’s. Additional market research utilizing the NASA SEWP V GWAC Provider Lookup tool revealed at least 10 SDVOSB contract holders that are authorized resellers of NVIDIA, Cisco, NetApp, and F5 and are capable of providing the complete BOM. Furthermore, the NASA SEWP V GWAC has a dynamic catalog where line items can be added based on customer requests; therefore, if the required item or service is within scope of the NASA SEWP V GWAC, items and services can be added to the catalog on a daily basis. As a result, there is a reasonable expectation for competition on the NASA SEWP V GWAC amongst SDVOSB resellers.

The CS conducted additional market research via the Small Business Administration Small Business Search on July 24, 2026, to identify certified SDVOSBs and Veteran Owned Small Businesses (VOSBs) that may be capable of fulfilling the aforementioned requirements. A search was conducted using North American Industry Classification System code 541519, “Other Computer Related Services.”

The search yielded 7,377 matches.

9. Other Facts:

None

Advocate for Competition Approval In my role as procuring activity Advocate for Competition, based on the foregoing justification, I hereby approve the acquisition of the brand name VISN 6 Biomed Virtualization, Server, Storage Expansion, and Application Delivery Refresh delivery order under the NASA SEWP V GWAC on an other than fair opportunity basis pursuant to the authority cited in Section 4 above, subject to availability of funds, and provided that the property and services herein described have otherwise been authorized for acquisition.