Attachment__1-_MINIMUM_REQUIREMENTS_v5___4-8-19_ks.docx

REVISED BY AMENDMENT 0001

MINIMUM REQUIREMENTS

Sensors and SONAR Systems Department, High Performance Computational Infrastructure (HPCI) Distributed Storage Cluster N66604-19-Q-0981

A distributed high performance network attached storage system is required to perform sensor data fusion, long term resilient, reliable, and scalable storage for advanced underwater systems. The key factors essential for the rack-mount distributed storage cluster consist of multiple homogenous servers/nodes with extremely fast network interconnects and managed by advanced software capable of providing end user network accessible distributed/clustered storage operations scaling up to several hundred petabytes (PB) in capacity. The minimum acceptable user accessible capacity of 640 terabytes with a raw capacity of 960 terabytes scaling no less than 8 nodes. The equipment must be new, fully operational, assembled, tested, and configured to include all associated cabling. Vendor must provide at a minimum 3 years of support for all hardware, software, and networking components associated with Software Defined Storage (SDS). Provide all associated components required in order to provide power, networking, and installation on standard rail mount system.

The table below lists the system minimum requirements for the distributed storage solution (per node):

Chassis: 4 RU(Max) Rack Mountable Form factor

Processor: 2x Current Generation Intel Xeon 10+ Cores Microprocessor @ 2.2+ GHZ

Memory: 128GB DDR4 Quad Channel with ECC (Error Correction Code)

Storage: Raw Capacity: 120 TB (total unformatted storage before file system configuration)

Accessible Capacity: 80 TB (usable space after file system redundancy/overhead)

Enterprise Class 15 x 8 TB SAS HDD | 1 x 3.2 TB SSD

Networking: Dual/Quad Network Interface Controllers capable of 10/25/40 gigabit Ethernet or

40 QSFP+ (physical network interface(s)) and must include all associated cabling.

Power Supply: 2 x Redundant 2KW Power Supply Units (PSU)

Accessories: Must include all associated components required in order to provide power, networking, and installation on standard rail mount system.

Ports Density: 32

Port Speed(s): 10/25/40 Gb Ethernet

Switch (Backend) 32 Ports, 2 x QSFP+ 40 Gb Ethernet (Network Interface Controller) 1U, 2PS

Switch (Frontend) 32 Port, 2 X QSFP+, 40 Gb Ethernet (Network Interface Controller) 1U, 2PS

Power Supply: 2 x Redundant 1KW Power Supply Units (PSU)

Accessories: Each switch must include all cabling required in order to provide power, networking, and installation on standard rail mount system.

Compatibility: Network connections between each node must be 2 x 40 Gb Ethernet. Frontend/Backend switches must be compatible with both Arista and Brocade network appliances for the purposes of integration and interoperability with existing infrastructure.

Software (Per Node)

Operating System: OS must be optimized specifically for tiered clustered and distributed storage allowing administrators to tailor resource allocation. Core OS platform must be Unix like (i.e. Linux, Unix, or BSD) and support continuous software upgrades and security patches. OS must support security enhancements such as DISA SRG’s/STIG’s and approved for use on DoD classified networks.

Clustered/Scalability: Each node must have specialized software installed which supports a clustered distributed SDS (Software Defined Storage) and be in a configured state. Software must support storage clusters consisting of many high performance, redundant, resilient, and scalable nodes capable of being managed from a centralized location scaling out to hundreds of petabytes.

File System Features: File system must implement a single namespace spanning multiple nodes each configured to utilize a parallel distributed networked file system for all data storage transactions while implementing erasure encoding, zero-copy, and journaling against metadata. Distributed File System must support the following capabilities or equivalents: High-performance (HP), Fully symmetric clustering (FSC), File striping across all nodes in a cluster, Automated software routines (ASR), Dynamic content balancing (DCB), Flexible data protection (DP), High availability (HA), Auto-balance, Scalable Rebuild, Erasure coding, virtual hot spare, and implement Web based/Command Line Administration tools.

Update/Upgrades/Management: Software must include the following features: Simultaneous/Rolling/Non-Disruptive upgrades, roll back capabilities and integration with vSAN (Virtual Storage Area Network), vSphere, and vCenter appliance transparently without the need for additional software. Software defined storage must support all of the following technologies for the purposes of ease of management: InsightIQ, SmartPools, SmartQuotas. SmartConnect, SnapShotIQ, Isilon for vCenter, Eyeclass DR Manager, SyncIQ, SmartLock, and SmartDedupe. Software must support minimal administrative efforts in order to install, upgrade, add/remove nodes, or conduct other general maintenance operations without interrupting storage availability or accessibility to clients. The features must support the following capabilities and/or equivalents.

Sharing/Networking Capabilities: Storage cluster software must be capable of providing either single or multiple namespace(s) which are accessible while using NFSv3 (Network File System Version 3.x), NFSv4 (Network File System Version 4.x), or SMB (Samba / CIFS) sharing protocols using Non-Disruptive Operations supporting both Windows and Linux client connections over well-known traditional network sharing protocols.

Security/Access Controls: Clustered storage must support traditional Microsoft Active Directory security groups against all files, folders, and objects stored within the distributed shared storage volumes. The software must support traditional ACL’s/SACL’s/DAC/RBAC using Active Directory users, groups, or a combination of the two against all files, folders, and objects stored within the distributed shard volumes in an Active Directory domain environment transparently (i.e. no third party applications are required on the clients).

Qualitative/Quantitative Evidence: Vendor must provide supporting artifacts in the form of benchmarks and documentation demonstrating performance metrics in either MB/s or GB/s with clients transferring data using both NFSv4 and SMB file sharing protocols.