HP Storage Architecture Design
Executive Summary
This document outlines the comprehensive storage architecture leveraging HP storage solutions for enterprise data management, including SAN, NAS, object storage, and data protection strategies.
1. Storage Architecture Overview
1.1 Storage Tiers
- Tier 0: NVMe All-Flash (HPE Primera) - Mission-critical applications
- Tier 1: SSD All-Flash (HPE Primera) - Production databases
- Tier 2: Hybrid Flash (HPE Nimble) - General purpose workloads
- Tier 3: Object Storage (HPE Apollo) - Unstructured data
- Tier 4: Archive (HPE StoreOnce) - Long-term retention
1.2 Storage Protocols
- Block Storage: Fibre Channel, iSCSI, NVMe-oF
- File Storage: NFS v4.2, SMB 3.1.1
- Object Storage: S3 API compatible
- Backup: Catalyst, NDMP, VTL
2. SAN Architecture
2.1 Fibre Channel SAN Design
Core SAN Switches
Production Fabric A:
├── HPE SN6650B 64Gb Director (Core-A1)
│ ├── 384 ports @ 64Gb
│ ├── Port Blade 1-6: Server connectivity
│ ├── Port Blade 7-8: Storage connectivity
│ └── Port Blade 9: ISL connectivity
└── HPE SN6650B 64Gb Director (Core-A2)
└── Redundant configuration
Production Fabric B:
├── HPE SN6650B 64Gb Director (Core-B1)
└── HPE SN6650B 64Gb Director (Core-B2)
└── Identical configuration to Fabric A
Edge SAN Switches
Rack-Level Connectivity:
├── HPE SN6600B 32Gb Switches (48-port)
├── Dual-fabric design (A/B)
├── 8x uplinks to core per switch
└── N_Port ID Virtualization (NPIV)
2.2 SAN Topology
Dual-Fabric Design:
├── Fabric A: Odd-numbered ports
├── Fabric B: Even-numbered ports
├── No ISL between fabrics
├── Core-Edge topology
└── Maximum 3 hops
2.3 Zoning Strategy
Zoning Configuration:
├── Type: Single Initiator Multiple Target (SIMT)
├── Method: PWWN-based soft zoning
├── Naming: SERVER_HBA#_ARRAY_PORT#
├── Peer Zoning: Enabled for array replication
└── Default Zone: Disabled
2.4 NVMe over Fabrics
NVMe-oF Configuration:
├── Transport: FC-NVMe
├── Arrays: HPE Primera with NVMe
├── Hosts: Gen10+ with NVMe support
├── Namespace: Dedicated per host
└── Multipath: ANA (Asymmetric Namespace Access)
3. HPE Primera Configuration
3.1 HPE Primera A670 Arrays
Primary Array (PRIMERA-01):
├── Controllers: 4-node (All-Active)
├── Cache: 1.536TB total
├── Drives: 48x 3.84TB NVMe SSD
├── Raw Capacity: 184TB
├── Usable Capacity: ~138TB (3+1 RAID)
├── Ports: 32x 32Gb FC, 16x 25GbE iSCSI
Secondary Array (PRIMERA-02):
└── Identical configuration
3.2 Primera Features
Data Services:
├── Thin Provisioning: Always enabled
├── Deduplication: Inline for VVs
├── Compression: Adaptive
├── Encryption: FIPS 140-2 Level 2
└── QoS: Priority Optimization
Availability Features:
├── Peer Persistence: Active-Active
├── Transparent Failover: Sub-second
├── Non-disruptive upgrades
└── 100% availability guarantee
3.3 Virtual Volume Layout
Production VVs:
├── PROD-DB-01 through 20: 10TB each
├── PROD-APP-01 through 30: 5TB each
├── PROD-WEB-01 through 40: 2TB each
└── Common Provisioning Groups (CPG):
├── CPG-SSD-R5: RAID 5 (3+1)
├── CPG-SSD-R6: RAID 6 (6+2)
└── CPG-SSD-R1: RAID 1 for logs
4. HPE Nimble Storage
4.1 Nimble Array Configuration
Nimble Arrays (4x HF40):
├── Controllers: Active-Standby
├── Cache: 768GB per array
├── SSD: 11.52TB per array
├── HDD: 120TB per array
├── Effective Capacity: 400TB+ per array
└── Connectivity: 4x 25GbE per controller
4.2 Nimble Pool Design
Storage Pools:
├── Pool-Performance:
│ ├── 100% SSD cache
│ ├── Tier 1 workloads
│ └── Sub-ms latency
├── Pool-Capacity:
│ ├── Hybrid configuration
│ ├── General workloads
│ └── 1-2ms latency
└── Pool-Archive:
├── Minimal SSD cache
├── Cold data
└── Best effort performance
4.3 Volume Collections
Application Groups:
├── VCG-VMware: All VMware datastores
├── VCG-Database: Database volumes
├── VCG-FileShare: File server volumes
├── VCG-Backup: Backup staging
└── Protection Schedules:
├── Hourly snapshots (24hr retention)
├── Daily snapshots (30 day retention)
└── Weekly snapshots (52 week retention)
5. NAS Architecture
5.1 File Services Design
HPE Nimble dHCI with Qumulo:
├── Nodes: 4x HPE ProLiant DL380
├── Software: Qumulo Core
├── Capacity: 800TB usable
├── Performance: 20GB/s throughput
└── Protocols: NFS v3/v4.2, SMB 3.x
5.2 File System Layout
Directory Structure:
/shares/
├── /home/ # User home directories
├── /dept/ # Department shares
├── /projects/ # Project collaboration
├── /applications/ # Application data
└── /archive/ # Archived file data
Quota Management:
├── User Quotas: 100GB default
├── Department: 10TB per dept
├── Projects: As requested
└── Soft/Hard limits enforced
5.3 Multi-Protocol Access
Access Methods:
├── SMB: Windows/Active Directory
├── NFS: Linux/Unix systems
├── S3: Object access gateway
├── REST API: Application integration
└── FTP/SFTP: Legacy support
6. Object Storage Architecture
6.1 HPE Scality RING
Object Storage Cluster:
├── Nodes: 12x HPE Apollo 4200
├── Drives: 24x 8TB per node
├── Raw Capacity: 2.3PB
├── Usable Capacity: 1.6PB (EC 8+3)
├── Connectors: S3, NFS, SMB, SWIFT
└── Durability: 14 nines
6.2 Bucket Strategy
Bucket Organization:
├── prod-webapp-assets: Static content
├── prod-backup-data: Backup archives
├── prod-archive-docs: Document archive
├── prod-media-files: Video/images
├── dev-test-data: Development data
└── Lifecycle Policies:
├── 30-day transition to IA
├── 90-day transition to Glacier
└── 7-year deletion policy
7. Data Protection Strategy
7.1 HPE StoreOnce Configuration
StoreOnce 5260 Systems:
├── Capacity: 1.7PB usable each
├── Deduplication: 20:1 average
├── Performance: 40TB/hr
├── Connectivity: 8x 25GbE
└── Features:
├── Catalyst stores
├── VTL for legacy
├── NAS backup targets
└── Cloud Bank Storage
7.2 Backup Architecture
Backup Tiers:
├── Tier 1: Local snapshots (instant)
├── Tier 2: StoreOnce (hours)
├── Tier 3: Replicated StoreOnce (daily)
└── Tier 4: Cloud archive (weekly)
Backup Software:
├── Primary: Veeam B&R v12
├── Database: Oracle RMAN
├── Secondary: HPE Data Protector
└── Cloud: Native cloud backup
7.3 Replication Strategy
Array-Based Replication:
├── Primera: Peer Persistence (Sync)
├── Primera: Remote Copy Async (15min)
├── Nimble: Volume replication (Hourly)
└── StoreOnce: Catalyst replication
Application-Based:
├── Database: Oracle Data Guard
├── VMware: vSphere Replication
├── Files: DFS-R / Rsync
└── Object: Cross-region replication
8. Storage Networking
8.1 IP Storage Network
iSCSI/NAS Network Design:
├── Switches: HPE 5940 (100GbE)
├── VLANs: Separate per protocol
├── MTU: 9000 (Jumbo frames)
├── Multipath: MPIO/ALUA
└── QoS: Storage priority class
8.2 Network Design
Storage VLANs:
├── VLAN 30: iSCSI-A (10.0.30.0/24)
├── VLAN 31: iSCSI-B (10.0.31.0/24)
├── VLAN 32: NFS (10.0.32.0/24)
├── VLAN 33: SMB (10.0.33.0/24)
├── VLAN 34: Replication (10.0.34.0/24)
└── VLAN 35: Backup (10.0.35.0/24)
9. Performance Optimization
9.1 Cache Optimization
Caching Strategy:
├── Read Cache: Predictive prefetch
├── Write Cache: Write-back with battery
├── SSD Cache: Auto-tiering enabled
├── Host Cache: Read cache on compute
└── Application: Redis/Memcached layer
9.2 I/O Optimization
Performance Tuning:
├── Queue Depth: 32-64 per LUN
├── Block Size: 64KB for databases
├── RAID Layout: Optimized per workload
├── Multipath: Round-robin
└── CPU Affinity: NUMA optimized
10. Capacity Management
10.1 Monitoring and Alerting
HPE InfoSight Integration:
├── Predictive analytics
├── Capacity forecasting
├── Performance bottlenecks
├── Proactive recommendations
└── Alert Integration:
├── ServiceNow tickets
├── Email notifications
├── SNMP traps
└── REST webhooks
10.2 Growth Planning
Capacity Triggers:
├── 80% utilization: Order planning
├── 85% utilization: Order placement
├── 90% utilization: Emergency expansion
└── Expansion Options:
├── Drive additions
├── Shelf expansions
├── Controller upgrades
└── Array additions
11. Security and Compliance
11.1 Data Security
Encryption:
├── Data at Rest: AES-256
├── Data in Flight: FC-SP, IPsec
├── Key Management: HPE ESKM
└── Compliance:
├── FIPS 140-2 Level 2
├── Common Criteria
├── PCI-DSS
└── HIPAA/HITECH
11.2 Access Control
Storage Access:
├── FC: PWWN authentication
├── iSCSI: CHAP/mutual CHAP
├── NFS: Kerberos v5
├── SMB: AD integration
└── Management: RBAC/MFA
12. Disaster Recovery
12.1 DR Storage Configuration
DR Site Storage:
├── HPE Primera A670 (Identical)
├── HPE Nimble HF40 (2 arrays)
├── HPE StoreOnce 5260
└── Reduced object storage
12.2 Recovery Objectives
Service Levels:
├── Tier 0: Zero RPO/RTO (Continuous)
├── Tier 1: 15min RPO, 1hr RTO
├── Tier 2: 1hr RPO, 4hr RTO
├── Tier 3: 24hr RPO, 24hr RTO
└── Test Schedule: Quarterly
13. Automation and Orchestration
13.1 Storage Automation
Automation Tools:
├── HPE OneView: Infrastructure as Code
├── Ansible: Playbook automation
├── PowerShell: Windows automation
├── REST APIs: Custom integration
└── Use Cases:
├── Volume provisioning
├── Snapshot scheduling
├── Replication management
└── Performance tuning
13.2 Self-Service Portal
Service Catalog:
├── Volume provisioning
├── Snapshot/restore
├── Performance class selection
├── Quota management
└── Integrated with ServiceNow
14. Storage Metrics and KPIs
14.1 Performance Metrics
- IOPS: Target 500K aggregate
- Throughput: Target 40GB/s
- Latency: <1ms for Tier 0/1
- Availability: 99.9999% target
14.2 Capacity Metrics
- Utilization: Target <80%
- Deduplication: 3:1 minimum
- Compression: 2:1 average
- Thin Provisioning: 2:1 oversubscription
15. Migration Strategy
15.1 Migration Approach
Migration Phases:
├── Phase 1: Non-critical data
├── Phase 2: Test/Dev systems
├── Phase 3: Production data
└── Phase 4: Mission-critical
Migration Tools:
├── HPE 3PAR Peer Motion
├── Storage vMotion
├── Host-based migration
└── Application-level migration
This comprehensive storage architecture provides enterprise-class data services with HP storage solutions, ensuring performance, availability, and scalability for all workloads.