SOO_FRD_Answers_151008.docx

W91RUS16GEF1

Statement of Objectives (SOO) and Functional Requirements Document (FRD) as of 151008

1. Would the Army please further explain the envisioned goal of "Removal of middleware"? Which specific types of middleware is the Army referring to? (2.3.3.1.h. Converged Architecture Goals)

A: The Army wishes to minimize layers in the application stack which creates a common interface for applications above it, and with respect to the lower level details of the infrastructure beneath it. Thus this is stated as a goal.

2. Is the Army referring to the provisioning of virtual resources (e.g., compute, storage, networking) for the requirement "Just-in-time hardware provisioning"? (2.3.3.1.j. Converged Architecture Goals)

A: Yes.

3. How many tenants does the Army envision for mission partners, organizational units, and other customer groups? (2.4.6 Multi-Tenancy)

A: There is envisioned to be 2-5 Major Tenancies with a possibility of more depending on the mission in addition there will subcommand on and installation

4. In addition to logical isolation of server and domain resources, does the Army envision a requirement for physical isolation of resources in separated GEF enclaves? If so, would the Army please elaborate on this requirement? (2.4.7 Security and Identity)

A: The Army believes that logical isolation should be sufficient.

5. Is the Army open to industry proposing alternative hardware and software components which represent current industry standards that are not currently on the DOD DISA APL at the time of contract award? ('8 - Site-Level Hardware, 9 - Site-Level Backup Software for Virtual Infrastructure, 10 - Enterprise System Hypervisor, and 11 - Enterprise System Management)

A: No, however systems that are currently in the process will be considered.

6. Would the Army please specify the particular Government security clearance levels that are required for vendor system support personnel? (12 Vendor Enterprise System's Support Defined Requirements)

A: SECRET for SIPRNET, UNCLAS no Foreign for NIPRNET.

Table 4. Requirements – Enterprise Hardware, Storage

7. 7.4.6 Can this requirement be stipulated a different way? If the platform meets all other requirements, is it important that it does so with multi-core or a highly parallel architecture? Additionally, if a platform is required to support Multi-Core processors, should it not be provided with them?

A: We are willing to consider solutions that are not based on multi-core technologies.

8. In 7.4.16 is the ‘user’ envisioned to be the storage administrator or the actual data end user of the service/solution. Are there specific data protection levels required?

A: The user is envisioned to be a storage manager. The protection levels are to be defined.

9. Re 7.4.19 the requirements lists 8/16Gb. Is this both or either? Additionally, although the switch in 7.3.5 lists 1GbE connectivity, there is no callout for that speed in this requirement. Is this an oversight?

A: 7.4.19 refers to the storage back end requirement. Para 7.3.5 refers to the requirement for Network connectivity. No oversight.

10. 7.4.30 Can you describe what data-at-rest encryption solutions are currently in use by the US Army?

A: Currently the Army uses Microsoft Bitlocker for Data-at-rest.

Table 5 Requirements – Site Hardware, File to Object-Object Gateway

11. Object Store/Cloud Gateway/Data Ingester requirements seem to be intermixed here. Can they be separated into distinct solutions for clarity as they provide differing functions?

For example, for each of these requirements, 8.5.1 through 8.5.36, does the requirements apply to the object store, ingester/gateway or both?

A: The requirement is stated in a way to allow industry to propose how these components will work together within their solution to meet the Army’s goal.

12. In 8.5.4 Please describe S3 as it relates to a “standard NAS feature”

A: The ability to integrate storage with Amazon’s S3 to extend native storage to the cloud.

13. In 8.5.5 Please describe latency expectations for sharing roaming profile access from all sites. Is a multiple copy strategy expected to reduce login latencies for truly mobile/roaming users?

A: Some latency would be expected but the Army would like to keep latency to a minimum. There is no expectation of one strategy over another. Please respond with your solution.

14. In 8.5.6 Can you better describe the global access topology? Many solutions run into locking and presence issues that increase network cross talk. A descriptive use case would be beneficial. Additionally, what would be the method for selection in the "selective pre population capabilities”?

A: Use Case; a user normally stationed at installation “x” and accesses file data on his local storage for his day to day job is given a short term assignment to go to installation “y”. This user must be able to gain access to his data in an efficient manner. It should appear to the user that the data is local. One method may be to prepopulate data at installation “y”.

15. 8.5.7. What proposed changes in metadata would trigger tiering?

A: The Army seeks the capability. The exact changes that may trigger the need for Tiering has not been defined. This could be included industry to propose it in there solution.

16. 8.5.9. "Supports replication, data protection and file restore without requiring additional HW or SW.” Replication of what? Replication of chunks using erasure encoding or similar is the object stores job. If at the higher level of the "global namespace and global locking” then replication does not apply because there is only one namespace that contains everything. In terms of data protection it is unclear if this is true data protection or just snapshots. Given the no additional HW or SW, is snapshotting what is actually meant as implied by the next sentence? For the snapshotting how many snapshots are required and going back how long?

A: The Army’s intent is to provide backup and restore capability within the solution. This includes the ability to point in time snapshots of specific file data as determined by users or administrators.

17. In 8.5.12 Please provide an example of a system wide Single point of failure that the requirement is meant to avoid.

A: Systems are expected to be redundant or deploy technology that accomplishes the same effect.

18. In 8.5.17 By replace Swift as Datastore, do you mean provide a Swift interface?

A: Yes.

19. In 8.5.18 Does the solution have to provide both Fixed-Block AND Variable-Block deduplication or can the solution provide one or the other type?

A: Either is acceptable.

20. In 8.5.20 Please clarify “no 0% downtime” A: System upgrades should be able to be performed during normal operations without impact to the Army’s ability to access the data from the device. It must appear to the user that there is 0% downtime.

21. 8.5.21. In the context of a file to object gateway what does the following mean, "Must support high-speed bus, disk and data protection architectures.”?

A: This requirement addresses its ability to integrate with the storage and data protection components of your solution.

22. 8.5.22. Is the intent to have 100% of data kept in all locations because tiering, caching, and selective pre population capabilities seem to conflict with the no additional latency requirement of "Any data can be accessed by any port within the system with no additional latency…”?

A: The intent is not to have 100% of the data kept at all locations but possibly regionalize data or storage to minimize latency. We are seeking best business practices

23. In 8.5.23 A central content repository is mentioned. Is it the intent to have a central repository across the enterprise, or central repositories at each of the DFMN sites?

A: The Army believes that the repositories would be at each DFMN site however, we seek recommendations from industry if there is a more efficient way of accomplishing our goals. The Army would consider a disturbed model.

24. In 8.5.26 What retention period is envisioned within which the user can perform a self-restore.

A: This still needs to be defined by the Army. We seek the capability and recommendations from industry.

25. In 8.5.29 Is the requested multi-site read/write operation expected to support multiple simultaneous users of the same data that is cached in accordance with 8.5.24?

A: Yes.

26. In 8.5.30 Can you describe this in more depth? E.g. what applications would you envision requiring this capability?

A: The Army desires the capability to maintain data integrity through this requirement. Enterprise NETOPS applications, there are no specific installation applications of concern at this time.

27. In 8.5.31 Is the intent of this requirement to allow existing backup solution to use this solution as a backup repository? If so, what are those existing backup solutions that would be expected to interface with this solution?

A: That is the intent. The Army has many varied backup solutions and understands that the solution may not be able to interface directly with all of them. We would ask industry to list the solutions they interface with.

28. Generally, what type of protection should be provided to the objects on the object store?

A: The Army desires that all objects be in at least two separate locations possibly separate geographical locations depending on the solution from industry.

Table 6 Requirements – Site Hardware, Backup to Disk

29. Image based backups are not always the most efficient or effective way to backup certain applications (notably large SQL databases). Should the backup also support agent architecture to better facilitate these types of backups?

A: The Army would be interested in reviewing any solution that provides greater efficiencies in these cases.

30. 9.6.5 Please describe source (this software?), in-flight (in flight across what media), and At Rest (what is the envisioned target for these backup streams and do these include the object storage requirement as outlined in 8.5.31)?

A: The initial question will depend heavily on the solution offered by industry however, these do include the requirements outlined in 8.5.31.

31. How many RFN sites are included in this procurement? Section 2.1 of the SOO mentions 11 sites, but Appendix B of the FRD lists 13 sites.

A: There are 11 RFN sites. The additional sites in Appendix B is for Army lab equipment to build out the Configuration Management portion of the solution.

32. What is driving the requirement for "to adequately manage and be managed by another hypervisor"?

A: The Army has not standardized on any one hypervisor however, we still have a requirement to centralize and simplify management of the hypervisor platforms. We are looking for industry recommendation on how managed a multiple hypervisor environment.

33. Does the Army envision the contractor to provide on-site day-to-day Operations & Maintenance (O&M) activities under this contract?

A: No.

34. Is the GMN merger into GEF envisioned in this first phase? or in subsequent phases?

A: The GMN will be merger into the SIPRNET requirements for DFN.

35. Is the government anticipating a single-award or multiple-award?

A: Single-award.

36. What is the anticipated period of performance?

A: The anticipated Period of Performance is to have base period and four option periods; subject to change

37. What requirement(s) are driving a separate procurement for IFNs?

A: Resources

38. Is it anticipated that the IFNs will utilize the same architecture as determined in this procurement?

A: The Army anticipates that the architecture will be compatible but no necessarily the same.

39. Is it anticipated that a draft RFP will be released?

A: A draft RFP is not anticipated; subject to change

40. 7.4.20 states that the solution must support block level storage. Is this the only storage type that will be considered? 5.1.3 states that NETCOM envisions a GEF topology that supports multiple types of storage.

A: Other solutions will be considered.

41. 7.4.24 states that de-duplication is required at the data level. Would that also include a requirement to perform de-duplication within the memory/cache tier for critical workloads?

A: Yes.

42. Table 8 refers to the Enterprise Manager Software. Is it a requirement for the hardware vendor to provide the ability for API’s that allow integration into the Enterprise manager? For example.. SCCM integration?

A: Yes. Provide solution in proposed solution

43. 12.9.1 calls for the Vendor to demonstrate patching capabilities. Is this testing part of the selection process? If not, where and how can this be accomplished?

A: Demonstration for this purpose can be in the form of documentation of how the patching would be accomplished through your solution.

44. 12.9.7 calls for vendors to test patching within the NETCOM Lab. Is NETCOM allowing vendors to provide the purposed equipment for testing? What is the process to enable the lab access?

A: This is a post award requirement.

45. 12.9.9 call for the Vendor to demonstrate the robustness of the architected solution reliability to allow for next business day support. How is NETCOM proposing this be accomplished?

A: Demonstration for this purpose can be in the form of documentation of how the support would be provided through your solution.

46. 12.9.16 calls for the Hardware solution to run STIG-ed approved software. Does this mean that any software including the Vendor supplied code/management tools/ software must be STIG-ed as well?

A: Yes, if applicable.

47. Does the Govt require non-disruptive scaling of the environment; i.e., the ability to add servers and storage with no service disruption?

A: Yes.

48. Does the Govt desire to minimize storage management by reducing the number of Data stores/Shares in the environment?

A: Yes.

49. Does the Govt require SNMPv3 support to enable the solution to send alerts to a centralized monitoring system? If not, what is the preferred reporting methodology?

A: Yes.

50. Does the Govt desire a single interface to provide non-disruptive upgrades for the entire infrastructure, including storage, servers, and hypervisor?

A: Yes.

51. Does the Govt desire the capability to rapidly STIG the Vendor environment and create a security-hardened environment through automated means?

A: Yes.

52. What is the Govt's plan to provide continuous monitoring of STIGs to ensure 0 x CAT1's and CAT2's of any chosen solution? Would it be beneficial if one of these responding vendors could provide an immediate and ongoing update process to achieve security standards as set forth by DISA?

A: The Army has many tools that monitor STIG compliance as well as inspections/audits of the systems. It would be very beneficial to have a process provided by vendors as part of their solution.

53. Will NETCOM award a new contract for this effort?

A: Yes.

54. What existing IDIQ Contract will be used for this effort if a new contract is not anticipated?

A: N/A.

55. Regarding the scope for the DFMN and RFN domains as defined on page 4 of the FRD – will these be awarded jointly or separately?

A: Anticipated to be awarded jointly; subject to change

56. Does the solution architecture for one domain (DFMN or RFN) need to be consistent with, or identical to, solution architecture in other domains?

A: The Army desires a standard solution if possible.

57. 7.4.20 states that the solution must support block level storage. Is this the only storage type that will be considered? 5.1.3 states that NETCOM envisions a GEF topology that supports multiple types of storage.

A: Other storage solutions will be considered.

58. 7.4.24 states that de-duplication is required at the data level. Would that also include a requirement to perform de-duplication within the memory/cache tier for critical workloads?

A: Yes.

59. Table 8 refers to the Enterprise Manager Software. Is it a requirement for the hardware vendor to provide the ability for API’s that allow integration into the Enterprise manager? For example.. SCCM integration?

A: Yes.

60. In Section 8 and in particular Table 5, in sections 8.5.1,14,19,24,34,35, and 36; reference is made to Object Store. Can you please provide a table, similar in detail to the other tables contained in the FRD, that sets the requirement(s) for the Object Store? Many Suppliers have products that may be classified as Object Stores, with very different definitions of what is meant by an Object Store.

A: The Army does not intend on detailing the Object Store any further. You solution should define object store as it applies to your technology.

61. 12.9.1 calls for the Vendor to demonstrate patching capabilities. Is this testing part of the selection process? If not, where and how can this be accomplished?

A: Demonstration for this purpose can be in the form of documentation of how the patching would be accomplished through your solution.

62. 12.9.7 calls for vendors to test patching within the NETCOM Lab. Is NETCOM allowing vendors to provide the purposed equipment for testing? What is the process to enable the lab access?

A: This is a post award requirement.

63. 12.9.9 call for the Vendor to demonstrate the robustness of the architected solution reliability to allow for next business day support. How is NETCOM proposing this be accomplished?

A: Demonstration for this purpose can be in the form of documentation of how the support would be provided through your solution.

64. 12.9.16 calls for the Hardware solution to run STIG-ed approved software. Does this mean that any software including the Vendor supplied code/management tools/ software must be STIG-ed as well?

A: Yes, if applicable.

65. Does the Govt require non-disruptive scaling of the environment; i.e., the ability to add servers and storage with no service disruption?

A: Yes.

66. Does the Govt desire to minimize storage management by reducing the number of Data stores/Shares in the environment?

A: Yes.

67. Does the Govt require SNMPv3 support to enable the solution to send alerts to a centralized monitoring system? If not, what is the preferred reporting methodology?

A: Yes.

68. Does the Govt desire a single interface to provide non-disruptive upgrades for the entire infrastructure, including storage, servers, and hypervisor?

A: Yes.

69. Does the Govt desire the capability to rapidly STIG the Vendor environment and create a security-hardened environment through automated means?

A: Yes.

70. What is the Govt's plan to provide continuous monitoring of STIGs to ensure 0 x CAT1's and CAT2's of any chosen solution? Would it be beneficial if one of these responding vendors could provide an immediate and ongoing update process to achieve security standards as set forth by DISA?

A: The Army has many tools that monitor STIG compliance as well as inspections/audits of the systems. It would be very beneficial to have a process provided by vendors as part of their solution.

Questions regarding support for the Overall Solution:

71. Do the hardware and software components have to be on the APL at time of submission, or at time of award? (FRD, page 9, Section 7) A: Time of award. However it needs to be stated in the proposal that the components are in the APL process.

72. Does the SLA of 4 hour acknowledgement and 24 hour resolution apply only to Category I (Severity I) issues? (FRD, Table 4, page 10) A: Yes.

73. Will the vendor need to adhere to a Government process and procedure for efficient patch maintenance, or will the vendor need to supply their own methodology (FRD, page 16, Table 9, requirement 12.9.1)?

A: Government process and procedures.

74. How will the Government make the determination of a “HazCon” as described on page 16 (FRD, page 16, Table 9, requirement 12.9.10)?

A: The Army determines “HazCon” as when redundancies have been exhausted and are no longer in place due to equipment or software failure.

75. Is it up to the Vendor to make a determination if a failure is imminent, or does the Government have parameters in mind to make this determination? It is important to make the distinction when addressing the support model, as it may impact the support plan. (FRD, page 16, Table 9, requirement 12.9.10) A: The Army will define the parameters based on the solution and recommendation from the vendor.

76. How many operators will need to be trained? Does it vary by location? (FRD, page 16, Table 9, requirement 12.9.11)?

A: Approximately 10-15 operators per Theater (CONUS, Europe, Pacific, Korea and South West Asia).

77. Which modes are acceptable for the training (e.g., instructor-led, virtual instructor-led, web based CBT, webinars, technotes, knowledgebase, etc.)? (FRD, page 16, Table 9, requirement 12.9.11)?

A: All modes of training are acceptable.

78. Will the Government require the vendor to perform the installation, testing and turn-up of the single integrated system once it arrives at the installation site? (FRD, page 16, Table 9, requirement 12.9.15)?

A: Yes.

79. Questions regarding the Statement of Objectives:

80. 2.2 PROGRAM MANAGEMENT: Will the Government please define the “industry best practices” that are referenced in Section 2.2? Additionally, and regarding the third program management focus point, will the Government publish the anticipated program schedule, technical performance, and risk documentation requirements that are expected from the contractor during performance?

A: The intent is for the vendor to propose the “industry best practices” as part of their solution based on their experience in deploying similar solutions. These additional elements are also expected to be proposed by the vendor.

81. 2.2.1 Integrated Product Teams: Will the Government further define the exact roles and responsibilities of the IPTs?

A: Yes, post award.

82. 2.3.3.1.b. & c. Does the Government anticipate requiring a particular service-level regarding end-to-end high-availability? For instance, is the contractor required to maintain availability at 98%?

A: No. SLA will be defined later

83. Ref.: SOO, 1.c., (sub-para 2 unnumbered) - Given that the IFNs are not a portion of the initial purchase, please provide information on what the current plan is to address the IFN requirements and provide the anticipated procurement actions and timeline for the IFNs?

A: The current approach to IFN’s is to re-purpose existing capacity where available and procure as needed when gaps are identified.

84. Ref.: SOO, 1.c. (sub-para 3 unnumbered) - Will the Army be providing specific information on the GMN infrastructure that will need to merged with the GEF?

A: GMN will be merged into SIPRNET requirements

85. Ref.: SOO, 2.3.2.,e. - Please define what the term ‘synchronization’ means as used in this sub paragraph.

A: The ability to provide consistent services through the JRSS architecture as it is implemented on GEF fielded sites.

86. Ref.: RFD - The requirements in Section 8 (Table 5) are quite comprehensive and contradictory for COTS Vendors to satisfy. It is recommended that the (storage) requirements for each particular functional area be broken out separately in order to best meet the Army’s objectives.

A: This is a recommendation, not a question. The Army does not intend to break out requirements any further. The Army wants industry to propose the best solutions based on the information provided.

87. Ref.: RFD, 10.7.1 - 4: This vendor suggests a Cloud Management Platform Can Cloud Management Platforms (CMP) that can manage multiple Hypervisors (including Hyper-V) be recommended?

A: Yes.

88. It is recommended that Hypervisors be managed by a distinct CMP and not (attempted to) be managed by other Hypervisors. This will reduce VM sprawl and unnecessary complexity.

A: Noted.

89. Ref.: RFD, 11.8.1: COMMENT - The functions described within this section are typically available with a Cloud Management Platform (CMP). Capabilities more advanced than mere provisioning of VMs can only be accomplished with a CMP. Functions such as automated provisioning that includes complex orchestration tasks, advanced performance monitoring, showback / chargeback, full lifecycle management are all available to a self-service customer via a Marketplace portal.

A: Noted.

90. Ref.: FRD Are there Onsite Support requirements and what levels?? Examples, 24/7/4hr, CTR (call to repair), Next business day, etc. This is a high availability project with a lot of back up and fail over to minimize downtime risk.

A: No Onsite support required other than installation. Other support is defined within the documents. The Army is open to other proposals.

91. Ref.: FRD Will DMR (Defective Media Retention) and or CDMR (Comprehensive Defective Material Retention) be required?

A: Yes.

92. Ref.: FRD Are specific Cleared Personnel required and what locations? – DoD and what level minimum?

A: Yes. All SIPRNET implementations require SECRET. Others require US personnel.

93. Ref.: FRD Will there be a requirement for US Soil, English speaking, help desk?

A: No.

94. Ref.: FRD Will the Government require full N + N for Power Redundancy?

A: No.

95. Ref.: RFD 8.5.7. Will the Government please clarify the intent of this requirement “Supports Adaptive Cloud Tiering – adapts to changes in file metadata, then automatically tiers corresponding content to the appropriate private or public cloud storage tier for enhancing TCO and security.”

A: The Army seeks the capability. The exact changes that may trigger the need for Tiering has not been defined.

96. Ref.: RFD 8.5.8. Will the Government please clarify the intent of this requirement “Multi-cloud support within same namespace, so to host different data sets in different cloud tiers; such as S3 enabled services.”

A: The intent is to have the capability to leverage the commercial cloud capabilities.

97. Ref.: SOO 1.2 b. Will the Government require the ability to automatically recover freed storage capacity without performance impact during production and an immediate return of capacity to the pool?

A: Yes.

98. Ref.: SOO 2.3.2 – Will the Government require the same OS across the small medium and large systems so as to simplify management and training?

A: That is preferred. The government is open to other proposed solutions.

99. Ref.: SOO 2.3.3b - Will the Government require four controllers in the SAN so as to provide high availability with high performance such that a controller failure will not cause write thru penalty?

A: The Army has a requirement for the solution to be high availability. We look to industry to define the specifics within their technology on how they will answer that requirement.

100. Ref.: SOO Will the Government require highly dense SSDs at minimum 3.84TB in a small form factor?

A: The solution must support SSD.

101. Ref.: SOO Will the Government require native file and block on the same SAN?

A: Yes.

102. Ref.: SOO Will the Government require the ability to replicate from a SAN without appliances front ending which would cause more latency and more management?

A: The Army requires the capability, appliances may be part of the solution.

103. Ref.: SOO Do all volumes/LUNs be presented to all controllers in a system and all controllers must be able to accept IOs from each other while system is running?

A: No.

104. Ref.: SOO Do RAID calculations need to be accomplished by an ASIC and not thru software since software would diminish performance and use controller CPU cycles?

A: The Army does not specify.

105. Ref.: SOO 2.3.4b – Will the Government require the solution to be able to be provisioned in minutes and scale quickly and easily?

A: The Army will consider a solution that had these characteristics.

106. Ref.: SOO 2.3.4C – Will the Government require 6 nines high availability?

A: The Army requires 4 nines. The Army open to any proposed solution meeting 4 nines or better.

107. Ref.: SOO -Will the Government require the ability to combine RAID types on the same spindles so as to increase capacity utilization and efficiency?

A: Yes.

108. Ref.: SOO 2.4.6 – Regarding Multi tenancy - must the SAN be required to provide secure administrative multi tenancy for up to 200 separate domains while maintaining massive parallel processing?

A: The Army will consider a solution that had these characteristics.

109. Ref.: FRD 7.0 The vendor recommends removal of this requirement. Not ALL H/W Product Categories are tested and certified for inclusion on the APL by vendor request. (APL Testing requires Government sponsorship). Will the Government provide the support and sponsorship required for applicable APL Certification, allowing additional time for testing and certifications?

A: No.

110. General. Q: Requirements for the DFWM are specified. Will there be detailed compute and storage specifications for the RFN & IFN? Detailed specification should include disk IOPS required for storage (not just capacity); CPU cores and/or total GHz required; and RAM required per DWFN/IFN/RFN A: Requirements are stated for both DFMN and RFN. (They are the same) as depicted in the FRD.

111. Section 5.1.1. Q: In this section the government refers to “Hardware Platform Independent” as a key requirement for the fabric pod. How do you anticipate adopting new technologies that may be hardware specific in this design? Will you drive towards lowest common denominator or favor and architecture that allows for new technology adoption as innovation occurs in industry?

A: The Army acknowledges that some aspects may be hardware dependent. However the Army intends to minimize this and strive for a “virtualization first” strategy. Industry proposal should address adopting technology innovation as it applies to the requirement.

112. Section 5.1.1. Q: In this section you specify 100% virtualized. Has the government considered the use of containers? Industry trends show strong evidence of value in using containers vs traditional virtualization tools, do you anticipate supporting containers in this architecture?

A: The Army will consider a solution that had these characteristics.

113. Section 5.1. Q: In this section it talks about overall convergence and COE, are you planning on converging the management tools as well?

A: The Army intends to converge and integrate as much as possible.

114. Section 5.1.1. Q: In this section it calls out hardware independence, what about being hypervisor independent as well (for both the architecture and management/automation tools)?

A: The Army will consider a solution that had these characteristics.

115. Section 5.1.3. Q: This section calls out a variety of storage protocols. Is there a plan to standardize on 1? If so, which one? If not, is there a technical reason to require support for so many?

A: There is no technical reason. The reality is that many storage investments have already been made by the Army and the intent is to leverage those investments wherever possible.

116. Section 5.1.3. Q: This section calls out both converged and hyperconverged solutions. Is there a reason to introduce different architectures in this design? With the goal of convergence and standardization, what drives the need for different converged solutions under the same architecture?

A: The Army’s intent is to get the best of both solution areas and to not limit competition.

117. Section 5.1.3, Figure 3. Q: In the figure it states that 4TB of hard drive space must be available on the servers. Since this is a highly scalable and available system can the government provide the rationale for requiring on-server storage when centralized storage is also requirement?

A) Is onboard storage required as a minimum specification?

A: Onboard storage is not required as a minimum specification.

118. Section 6.2. Q: The government is requiring elasticity and the appearance of infinite capacity. Is the government actually requiring an infrastructure-as-a-service offering where the vendor is completely responsible for the entire lifecycle, growing, shrinking, and maintenance of the infrastructure while the government simply consumes what they require and only pay for what they consume? This aaS approach seems to follow the requirements in section 12.

A: The intent is for the vendor to assist the Army in providing an Army hosted aaS to their customers. The intent is not for the vendor to provide this service in this phase. The Army is responsible for the on boarding process, the on-boarding requirements are not included in this SOO.

119. Section 6.4. Q: This section specifies a need for Predictability but provides no method to monitor/measure services performance. Should vendors propose solutions to provide these capabilities as well?

A: Yes.

120. Section 6.6. Q: This section calls out for multi-tenancy. What level of security is required? PL3 or Pl4 as an example?

A: 3PL.

121. Section 7, General. Q: In this section it does not refer to local HDD, but Figure 3 shows a 4TB local requirement. Typically rack servers have some local storage, while blades do not to take advantage of the SAN. Can you provide more details about the need for local storage (if needed at all) and if needed, how it will be used?

A: Onboard storage is not required as a minimum specification.

122. Section 7, First Paragraph. Q: Is inclusion on the DOD DISA APL a requirement or is an Army CoN sufficient. Alternate multi-purpose solutions that combine capabilities such as data & storage acceleration may not have an applicable classification on the APL.

A: Yes.

123. Section 7.2.1-7.2.7 Q: Many requirements (core count for example) are very dated. Intel currently provides up to 18-core per socket. And memory runs at different speeds (up to 2133). Can you provide more specifics to ensure the best design?

A: The specifications provided were intended to give a general scope of the Army’s requirement. We encourage vendors to propose more effective and efficient ways on meeting or exceeding these minimums.

124. Section 7.3.2, Table 3. Q: In this section it states: "Must have local-to-host virtualized storage controllers for supporting IP Storage”. Can you be more specific about this requirement for IP based storage?

A: If the storage solution chosen does in fact use 10GbE network connectivity that would require a networking solution that could support IP based storage. Additional solutions will be considered.

125. Section 7.3.4, Table 3. Q: will you accept 40Gb ports as well?

A: Yes.

126. Section 7.3.10. Q: Out of Band (OOB)/Management networks are typically virtually separate without a completely separate infrastructure (costs of rack space, cabling, cooling, etc) in the Data Center space. Cisco shows a best practice here using VLANs (http://www.cisco.com/c/en/us/td/docs/solutions/Enterprise/Security/SAFE_RG/SAFE_rg/chap9.html#wp1054536). Is there a need for a completely separate infrastructure (including dedicated WAN circuits, etc) is simply providing a virtually separate (VLANs, VRFs, etc) sufficient?

A: Virtual segregated is sufficient. The government would consider other solutions Section 7.4.6. Q: Can this requirement be stipulated a different way? That is, if the platform meets all other requirements, is it important that it does so with multi-core or a highly parallel architecture? A)

a. Additionally, if a platform is required to support Multi-Core processors, should it not be provided with them?

A: We are willing to consider solutions that are not based on multi-core technologies.

127. Section 7.4.16. Q: In this section is the ‘user’ envisioned to be the storage administrator or the actual data end user of the service/solution?

a. Are there specific data protection levels required?

A: There will be a separate storage Administrator but, we expect some capabilities to be available to the user.

128. Section 7.4.30. Q: Can you describe what data-at-rest encryption solutions are currently in use by the US Army as it relates to this solution’s requirements?

A: The Army has deployed Bitlocker for Data-at-rest.

129. Section 8, General. Q: Object Store/Cloud Gateway/Data Ingester requirements seem to be intermixed here. Can they be separated into distinct solutions for clarity as they provide differing functions? As an example, for each of these requirements, 8.5.1 through 8.5.36, does the requirements apply to the object store, ingester/gateway or both?

A: The requirement is stated in a way to allow industry to propose how these components will work together within their solution to meet the Army’s goal.

130. Section 8.5.2. Q: This section references drive encryption. As the SAN allows for true application and server mobility (from a stateless compute perspective) is the encryption requirement SAN specific? Or is there a need for local encrypted storage that preempts the ability to deliver a stateless system?

A: SAN specific for data-at-rest.

131. Section 8.5.4. Q: Can the government please describe S3 as it relates to a “standard NAS feature” A: The ability to integrate storage with Amazon’s S3 to extend native storage to the cloud.

132. Section 8.5.5. Q: Can the government please describe latency expectations for sharing roaming profile access from all sites. Is a multiple copy strategy expected to reduce login latencies for truly mobile/roaming users?

A: Some latency would be expected but the Army would like to keep latency to a minimum. There is no expectation of one strategy over another. Please respond with your solution.

133. Section 8.5.6. Q: Can the government better describe the global access topology? Many solutions run into locking and presence issues that increase network cross talk. A descriptive use case would be beneficial.

a. What would be the method for selection in the "selective pre population capabilities”?

A: Use Case; a user normally stationed at installation “x” and accesses file data on his local storage for his day to day job is given a short term assignment to go to installation “y”. This user must be able to gain access to his data in an efficient manner. It should appear to the user that the data is local. One method may be to prepopulate data at installation “y”.

134. Section 8.5.7. Q: What proposed changes in metadata would trigger tiering?

A: The Army seeks the capability. The exact changes that may trigger the need for Tiering has not been defined.

135. Section 8.5.9. Q: In this section it states: "Supports replication, data protection and file restore without requiring additional HW or SW.” Can the government clarify the replication of what?

Replication of chunks using erasure encoding or similar is the object stores job. If at the higher level of the "global namespace and global locking” then replication does not apply because there is only one namespace that contains everything. In terms of data protection it is unclear if this is true data protection or just snapshots. Given the no additional HW or SW, is snapshotting what is actually meant as implied by the next sentence?

For the snapshotting how many snapshots are required and going back how long?

A: The Army’s intent is to provide backup and restore capability within the solution. This includes the ability to point in time snapshots of specific file data as determined by users or administrators.

136. Section 8.5.12. Q: Can the government please provide an example of a system wide Single point of failure that the requirement is meant to avoid.

A: The Army does not want the failure of one component to cause system failure.

137. Section 8.5.17 . Q: Can the government clarify what is meant by replace Swift as Datastore, do you mean provide a Swift interface?

A: Yes.

138. Section 8.5.18. Q: Does the solution have to provide both Fixed-Block AND Variable-Block deduplication or can the solution provide one or the other type?

A: Either is acceptable.

139. Section 8.5.20. Q: Can the government please clarify “no 0% downtime” A: System upgrades should be able to be performed during normal operations without impact to the Army’s ability to access the data from the device. It must appear to the user that there is 0% downtime.

140. Section 8.5.21. Q: In the context of a file to object gateway what does the following mean, "Must support high-speed bus, disk and data protection architectures.”?

A: This requirement addresses its ability to integrate with the storage and data protection components of your solution.

141. Section 8.5.22. Q: Is the intent to have 100% of data kept in all locations because tiering, caching, and selective pre population capabilities seem to conflict with the no additional latency requirement of "Any data can be accessed by any port within the system with no additional latency…”?

A: The intent is not to have 100% of the data kept at all locations but possibly regionalize data or storage to minimize latency. The Government would consider a distribute model.

142. Section 8.5.23. Q: A central content repository is mentioned. Is it the intent to have a central repository across the enterprise, or central repositories at each of the DFMN sites?

A: The Army believes that the repositories would be at each DFMN site however, we seek recommendations from industry if there is a more efficient way of accomplishing our goals. The Government would consider a distribute model.

143. Section 8.5.26. Q: What retention period is envisioned and required within which the user can perform a self-restore?

A: This still needs to be defined by the Army. We seek the capability and recommendations from industry.

144. Section 8.5.29. Q: Is the requested multi-site read/write operation expected to support multiple simultaneous users of the same data that is cached in accordance with 8.5.24?

A: Yes

145. Section 8.5.29. Q: Can the government describe this in more depth? That is, what applications would you envision requiring this capability?

A: The Army desires the capability to maintain data integrity through this requirement. There are no specific applications of concern at this time. Reference should be 8.5.30.

146. Section 8.5.31. Q: Is the intent of this requirement to allow existing backup solution to use this solution as a backup repository? If so, what are those existing backup solutions that would be expected to interface with this solution?

A: That is the intent. The Army has many varied backup solutions and understands that the solution may not be able to interface directly with all of them. We would ask industry to list the solutions they interface with.

147. Section 8, General. Q: What type of protection should be provided to the objects on the object store?

A: The Army desires that all objects be in at least two separate locations possibly separate geographical locations depending on the solution from industry.

148. Section 9, Table 6. Q: Image based backups are not always the most efficient or effective way to backup certain applications (notably large SQL databases). Should the backup solution also support agent architecture to better facilitate these types of backups?

Should the solution support both options to enable flexibility to provide the best service to each backup scenario?

A: The Army would be interested in reviewing any solution that provides greater efficiencies in these cases.

149. Section 9.6.5. Q: Can the government please describe source (this software?), in-flight (in flight across what media), and At Rest (what is the envisioned target for these backup streams and do these include the object storage requirement as outlined in 8.5.31?

A: The initial question will depend heavily on the solution offered by industry however, these do include the requirements outlined in 8.5.31.

150. Section 11.8.4, Table 8. Q: This section asks for a single pane of glass (SPOG) for Big Data support. Will that SPOG requirement be coupled with the SPOG requirement or is this another tool?

A: The Army currently envisions this as a separate tool however there is no reason that it has to be.

151. Sections 11.8.15 through 11.8.20, Table 8. Q: There is much discussion about SDN and support for VM’s in the virtual network. The only reference to underlying physical networks is 11.8.20 and “self-healing”. Is there a requirement for the Management Software to interact, program and represent the status of the underlying physical network? Also, is there a requirement to map VM traffic to the underlying physical network? If not, how will the government ensure that the network is “self-healing” in a satisfactory way?

A: The Army has several tools deployed to monitor the health of the network in our current deployments. Define how self-healing is accomplished within your solution.

152. General / Section 11.8.17-19. Q: SDN capabilities are specified as a requirement but again no method to monitor and troubleshoot SDN exists within this architecture – should vendors propose solutions to address this needed component?

A: Yes.

153. Section 12. Table 9. Q: The requirements in this section seem to indicate that government will not manage and maintain the system. Is it the intent of the government to procure a solution where the government simply consume resources and is billed for the resource usage?

A: The intent is for the vendor to assist the Army in providing an Army hosted aaS to their customers. The intent is not for the vendor to provide this service in this phase.

154. Section 12. it in this section that “The vendor must have personnel with the appropriate technical expertise as well as Government security clearances for NETCOM to effectively project manage the vendor’s build, delivery, installation, maintenance, and decommissioning of the hardware and software components of the fabric pod. “ Can the government clarify if this means that the equipment will be fully maintained and operated by the offeror?

A: The intent is for the vendor to assist the Army in providing an Army hosted aaS to their customers. The intent is not for the vendor to provide this service in this phase. The intent is the vendor to support technology for duration of the maintenance agreement.

155. Can the government clarify if this means that a separate support contract to augment the O&M organization (e.g. RCCs, NETCOM, etc) will be procured to manage the system A: The intent is for the vendor to assist the Army in providing an Army hosted aaS to their customers. The intent is not for the vendor to provide this service in this phase.

156. If it is the intent of the government to procure support, maintenance, and management of the system as part of the hardware and software procurement, can the government provide a more detail:

Detailed PWS Expected Period of Performance Expected place of performance (or done through the offeror’s NOC) Type of contract: FFP, CPFF, T&M Detailed scope of Section 12.9.12 (how many, how much data, etc)Section 12.9.15. Q: Can the government clarify the following regarding this section:

A: The intent is for the vendor to assist the Army in providing an Army hosted aaS to their customers. The intent is not for the vendor to provide this service in this phase.

157. Shall the equipment be simply shipped in a single shipment or shall be configured prior to shipment (in a single shipment) to each site?

A: The Army would prefer the equipment to be configured prior to shipment.

158. Shall the equipment be shipped in a rack that the equipment has been fully cabled and all equipment labeled OR simply shipped in a single shipment, palletized?

A: Please respond as appropriate for your solution.

159. Will the offeror be responsible for on-site physical installation and/or on-site configuration of equipment for administrators to be able to remotely access the devices? Or will this be the responsibility of the local NEC and the RCC?

A: Offerer will be responsible for installation and configuration.

160. Section 13.6 and 6.7. Q: These sections refer to security at a high level. What is the government’s position on preventing breaches in security from a rogue system? Malware protection and auditing? How will the government address insider threat in this solution?

A: These capabilities are important to the Army but are not specifically addressed in this acquisition.

161. Section 1.1 describes ‘roll-up’ capabilities for visibility/management. Is it the Army’s intent or desire to be able to rapidly shift hosting workloads between nodes at different locations? Is the vision to provide a ‘Tiered’ data architecture between nodes for backup purposes?

A: Yes.

162. Section 1.1.c describes hosting ‘local’ applications on IFNs. Will NETCOM employ the same definition for local vs enterprise applications as the Army CIO/G6 previously published in their Core Data Center policy? If so, can the Army describe how they will determine which specific host will be used for a particular application?

A: For this phase the definition of Enterprise applications is limited to those described in 2.5 of the SOO.

163. Section 1.3 depicts several components/services that can be fairly easily correlated to a specific workload for a virtual hosting platform. However, some such as ‘IT Asset Management,’ ‘Unified Capability,’ and ‘Enterprise Support Management System’ are more ambiguous. Does the Army have a specific list of applications that they intend to host from each node/tier in the GEF?

A: The Army does not have a complete list for each node/tier of the GEF other than the applications described in 2.5 of the SOO.

164. Section 2.1 says that the final architecture will be informed by industry. Is the Army going to pursue an architecture development task as part of the GEF, or separately?

A: No. Your proposal should include all required documentations

165. Section 2.2 briefly refers to using available Government Furnished Equipment (GFE) to reduce costs. Can the Army provide a detailed list of hardware resources that may be available for use?

A: No, this applies to the implementation of IFNs in a later phase.

166. Is Storage Area Network (SAN) hardware a firm requirement, or will the Army accept integrated storage to meet the hosting requirements?

A: Integrated storage will be considered.

167. Does the Army have a data archiving schema or architecture developed, and is there interest in consolidating data at particular nodes/tiers within the GEF to reduce costs?

A: Not during this phase, but during the IFN phase it will be required..

168. The move to a private cloud hosting environment requires adherence to standardized services and methods.