We’re evaluating migration from on-premise CloudSuite to cloud deployment and need to understand real-world performance implications for our revenue management workload. We process approximately 500,000 revenue transactions monthly with complex recognition rules.
Our on-prem environment is highly tuned with dedicated database servers and we’re concerned about performance in a multi-tenant cloud environment. Specifically worried about month-end revenue recognition processing which currently takes 6-8 hours and batch posting operations.
For those who’ve migrated high-volume revenue workloads to cloud, what performance differences did you experience? How do cloud and on-prem compare for scaling and latency concerns during peak processing periods?
We migrated a similar volume last year. Surprisingly, our cloud performance exceeded on-prem after optimization. The key was working with Infor to right-size our cloud instance. They provide performance tiers that can match or exceed typical on-prem setups. Our month-end processing actually dropped from 8 hours to 5 hours, mainly because cloud infrastructure is newer and faster.
Having completed this migration journey with high-volume revenue processing, I can provide comprehensive insights on the performance comparison.
High-Volume Revenue Workload Performance: The critical factor is understanding what drives your current on-prem performance. If your 6-8 hour processing window is primarily CPU-bound (complex revenue recognition calculations), cloud performance will be comparable or better because Infor uses current-generation processors. If it’s I/O-bound (database writes and reads), performance depends on your cloud tier selection. We found that revenue recognition processing - which is computation-heavy - actually ran 15-20% faster in cloud due to better CPU performance. However, batch posting operations with high database transaction rates initially ran slower until we upgraded to a higher IOPS tier for our cloud database.
Cloud Versus On-Prem Architecture: The fundamental difference is resource allocation flexibility. On-prem gives you dedicated hardware but fixed capacity. Cloud provides elastic scaling but shared infrastructure. For revenue workloads with predictable monthly peaks, cloud’s ability to temporarily scale resources during month-end processing is valuable. We now request additional compute resources for the 3-day month-end window and scale back afterward. This would require permanent hardware investment on-prem. The multi-tenant concern is largely mitigated by Infor’s resource isolation - your workload runs in dedicated containers with guaranteed resource allocations specified in your cloud tier.
Scaling and Latency During Peak Periods: This is where cloud architecture shows distinct advantages. During month-end processing, we can increase parallel processing threads without worrying about impacting other operations because cloud infrastructure auto-scales supporting services. Latency for batch jobs is actually lower in cloud because Infor’s data centers have high-speed internal networking. For interactive user operations during peak periods, latency is more about geographic proximity than cloud versus on-prem. Users within 500 miles of the data center experience 20-40ms latency. International users see 100-200ms, which is noticeable but not problematic for most revenue management tasks.
Migration Performance Strategy: When planning your migration, request a cloud sizing assessment from Infor based on your transaction volumes. They’ll analyze your current on-prem resource utilization and recommend appropriate cloud tiers. Budget for one tier higher than their baseline recommendation - the incremental cost is minimal compared to performance risk. Implement phased migration with parallel processing validation. Run your month-end close simultaneously in both environments for 2-3 cycles to validate performance before cutting over completely. This parallel approach identified several query optimization opportunities that improved both environments.
Real-World Outcome: Our revenue processing performance in cloud now matches or exceeds our highly-tuned on-prem environment, with the added benefits of elastic scaling, automated infrastructure management, and faster disaster recovery. The 10-15% performance variance we initially experienced disappeared after proper cloud configuration tuning.
Interesting that cloud performed better. Were there any specific optimizations you made post-migration? Also, how does network latency affect user experience for revenue posting and inquiry screens?
Don’t overlook the infrastructure management benefits. On-prem requires you to maintain hardware, apply patches, manage backups, and handle disaster recovery. Cloud shifts all of that to Infor. Our IT team freed up 30% capacity by eliminating on-prem infrastructure management. That operational efficiency often outweighs minor performance differences.
The multi-tenant concern is valid but overblown. Infor uses resource isolation in their cloud architecture. During our migration, we ran parallel processing for three months - same workload on-prem and cloud. Cloud was consistently within 10-15% performance of on-prem, and that gap closed as we tuned our cloud configuration. The bigger benefit is scaling - cloud can burst capacity during month-end without capital investment in hardware.
Latency is the real consideration. For batch processing, it’s negligible. For interactive screens, expect 50-150ms additional latency depending on your location relative to Infor’s data centers. Users in the same region as the data center barely notice. International users see more impact. We implemented regional caching for frequently accessed reference data which helped significantly.
Post-migration optimizations included adjusting batch job schedules to leverage cloud auto-scaling, implementing parallel processing for revenue recognition batches, and optimizing our SQL queries for cloud database configuration. Infor’s cloud team provided query performance analysis that identified several inefficient queries we’d been running for years on-prem without realizing the impact.