PostgreSQL Performance Analysis: HWA 10.2.4 vs 10.2.3 Comprehensive Comparison

​This analysis presents the first comprehensive performance comparison of HWA (Workload Automation) running on PostgreSQL database, comparing versions 10.2.3 and 10.2.4 to demonstrate the evolution and optimization achieved in the latest release.

Executive Summary
- HWA 10.2.4 with PostgreSQL shows significant improvements in scheduling performance
- Enhanced workload handling capability with 43% increase in job volume (400K vs 280K jobs)
- Improved response times during peak loads while maintaining consistent throughput
- Stable critical operations performance across both versions

Test Environment Configuration
Both tests maintained consistent infrastructure for reliable comparison:

Common Configuration:
- Infrastructure: VMware 6 + RHEL Server 9.4
- Workload Distribution: Large PVT daily production plan
- Peak Load Pattern: 10K jobs scheduled in 10-minute bursts
- Extended Load: 600 jobs/min sustained workload
- Workstations: 17 workstations, 3,200+ job streams

Version-Specific Details:
- 10.2.4: Build 20240413, PostgreSQL 16.4, WebSphere Liberty 25.0.0.3, 400K+ jobs
- 10.2.3: Build 20241120, PostgreSQL 16.4, WebSphere Liberty 24.0.0.9, 280K+ jobs

---

Performance Metrics Comparison

🚀 Peak Load Performance Analysis

⚡ Extended Load Performance (18:00-23:00)

🔧 Critical Operations Performance

Performance Analysis Deep Dive

Major Improvements in 10.2.4
🎯 Enhanced Job Processing Efficiency: Despite handling 43% more jobs (400K vs 280K), HWA 10.2.4 demonstrates:

• 31% faster schedule delays during peak loads (18s vs 26s)
• 44% improvement in mirroring delays during peak loads (20s vs 36s)
• Consistent throughput of 953 jobs/min maintained despite increased workload
  
  

⚡ Optimized Response Times: The significant reduction in both scheduling and mirroring delays indicates substantial optimization in the core workload automation engine.

🔧 Stable Critical Operations: Both MAKEPLAN and CHECKSYNC operations maintain identical performance (28s and 65s respectively), demonstrating stable foundational operations.

Resource Utilization Trade-offs

📈 Expected Resource Increase: The higher resource utilization in 10.2.4 is proportional and exceptionally well-optimized:

• 43% more jobs processed with only 11% CPU increase (MDM), 6% CPU increase (PostgreSQL), 9% disk increase (MDM), and 8% disk increase (PostgreSQL)
  
• Resource efficiency ratio dramatically improved: significantly more work with minimal resource increases across all metrics
  
• All resource utilization increases are very reasonable given the substantially larger workload

💡 Efficiency Gains: When normalized for workload size, 10.2.4 shows exceptional resource efficiency:

• Jobs per CPU unit: 10.2.4 processes significantly more jobs per percentage of CPU used
  
• Jobs per disk I/O unit: Improved I/O efficiency with more jobs processed per disk utilization percentage
  
• Response time improvement: Faster processing despite substantially larger workload

---

PostgreSQL-Specific Insights

Database Performance Characteristics
​
🗃️ PostgreSQL 16.4 Platform Stability: Both versions running on the same PostgreSQL 16.4 platform demonstrate:

• Pure HWA optimization effects without database platform variables
  
• Consistent database performance baseline across both tests
  
• Application-level improvements clearly isolated and measurable
  
• PostgreSQL 16.4 capabilities fully utilized in both versions
  
  

Version Compatibility
📦 PostgreSQL Version Consistency:

• 10.2.3: PostgreSQL 16.4
  
• 10.2.4: PostgreSQL 16.4 (same version)
  
• Consistent database platform isolating HWA application improvements
  
• Pure application optimization effects clearly demonstrated

---

Recommendations

For Production Deployment
✅ Strongly Recommended: HWA 10.2.4 with PostgreSQL for:

• High-volume production environments requiring 400K+ daily jobs
• Performance-critical workloads needing fast scheduling responses
• PostgreSQL-based infrastructures seeking optimal integration

Capacity Planning
📊 Resource Planning Guidelines:

• CPU: Plan for ~41% MDM CPU and ~26% PostgreSQL CPU for 400K+ job workloads
• Disk I/O: Ensure sufficient I/O capacity for ~64% MDM and ~70% PostgreSQL disk utilization
• Throughput: Expect consistent 953 jobs/min peak throughput capability
  
  

Migration Strategy
🔄 Upgrade Path:

• Pure HWA application upgrade from 10.2.3 to 10.2.4 with consistent PostgreSQL 16.4
  
• Database platform stability ensuring performance comparisons reflect only application improvements
  
• Clear optimization benefits from HWA application enhancements

---

Conclusion

HWA 10.2.4 with PostgreSQL delivers exceptional performance improvements while handling significantly larger workloads. The 31% improvement in schedule delays and 44% improvement in mirroring delays during peak loads, combined with the ability to process 43% more jobs, makes it the clear choice for enterprise PostgreSQL deployments.

The significant performance improvements can be attributed directly to HWA 10.2.4 application optimizations, with both versions running on the same PostgreSQL 16.4 platform. This provides a clear demonstration of the application-level enhancements without database platform variables.
​
PostgreSQL 16.4 serves as an excellent stable database backend for both HWA versions, with the performance gains in 10.2.4 clearly demonstrating the superior application optimizations and enhanced scaling characteristics for enterprise workload automation.

---

​Analysis based on comprehensive testing with 280K-400K job workloads on consistent PostgreSQL 16.4 platform

​Alessandro Tomasi
Senior Technical Lead
Alessandro is a Software Engineer. He has an academic background in Computer Engineering,
in particular in technology customer support and automation testing

---

Marco Gangemi
Software Engineer
Marco comes from an academic background in Physics. He joined HCL in 2018.
​Since then, he has specialized in DevSecOps activities. He’s recently started on Performance tests.

---