# Architecture selection
<a name="architecture-selection"></a>


|  LSPERF01: How do you select architectural patterns that accommodate the computing needs of genomic sequencing and molecular modeling?  | 
| --- | 
|   | 

 When selecting architectures for genomic sequencing and molecular modeling, prioritize scalable compute patterns that efficiently handle variable workloads. Evaluate high-performance computing options with high memory-to-CPU ratios and GPU acceleration to optimize performance while maintaining data integrity for your scientific workflows. 


|  LSPERF02: How do you manage diverse datasets ranging from basic records to terabyte-scale scientific datasets including sensitive data?  | 
| --- | 
|   | 

 Implement tiered storage strategies using hot to cold storage tiers for cost-efficient management of large scientific datasets while keeping sensitive patient records in encrypted database instances. Deploy auto scaling data processing pipelines that efficiently handle both high-velocity instrument data and carefully controlled clinical information flows. 


|  LSPERF03: How do you prioritize optimization between research computing and clinical application components in your architecture?  | 
| --- | 
|   | 

 Analyze workload characteristics to identify optimization targets. Deploy high-performance computing resources for research pipelines requiring massive parallel processing, while prioritizing availability and consistent performance for clinical applications. Use dedicated and isolated environments with tailored service configurations to meet distinct research and clinical requirements. 


|  LSPERF04: How do you approach performance considerations in your designs for long-term clinical trials and longitudinal studies?  | 
| --- | 
|   | 

 For long-term clinical trials, design architectures that maintain consistent performance over extended timeframes. Implement data lifecycle policies that optimize storage costs while preserving query performance. Deploy versioned infrastructure-as-code to create reproducible environments that can be efficiently recreated years later for regulatory adherence. 

**Topics**
+ [LSPERF01-BP01 Design and benchmark computing architecture for genomic workloads to optimize cost-performance ratio](lsperf01-bp01.md)
+ [LSPERF01-BP02 Specialized hardware selection and optimization for genomic and molecular workloads](lsperf01-bp02.md)
+ [LSPERF01-BP03 Performance optimizations should validate data integrity](lsperf01-bp03.md)
+ [LSPERF02-BP01 Data-aware storage tiering](lsperf02-bp01.md)
+ [LSPERF02-BP02 Secure data separation by classification](lsperf02-bp02.md)
+ [LSPERF02-BP03 Elastic data processing pipelines](lsperf02-bp03.md)
+ [LSPERF03-BP01 Workload-specific performance analysis](lsperf03-bp01.md)
+ [LSPERF03-BP02 Environment isolation by workload type](lsperf03-bp02.md)
+ [LSPERF03-BP03 Tailored service configuration by use case](lsperf03-bp03.md)
+ [LSPERF04-BP01 Performance consistency through clinical trial lifetime](lsperf04-bp01.md)