Performance Tuning
Performance Tuning Lab: Workload Alignment
Align queue depths, LUN layouts, and host multipathing without chasing mythical one-click fixes.
- Duration
- 5 days lab-heavy
- Format
- In-person lab
- Tuition (informational)
- ₩2,100,000
Overview
Performance work belongs in the daylight of measured change. You profile staged workloads, compare before-and-after traces, and learn when to stop tuning. We highlight procurement-ready conversations: what to ask vendors for, what to prove internally first.
What is included
- ● Trace reading drills with annotated baselines
- ● Queue depth experiments with guardrails
- ● Host multipathing lab on Linux and Windows samples
- ● Vendor-neutral IO pattern library
- ● Mentor review of your tuning memo
- ● Capstone presentation to the cohort
Outcomes
- Write a tuning memo with evidence, scope, and rollback.
- Pick two metrics that actually matter for your workload.
- Say no to vanity tuning with diplomatic language.
MP
Lead mentor
Minseo Park
Lead SAN instructor focused on performance narratives operations trust.
Participant questions
Do I bring my own workloads?
Optional anonymized traces are welcome for office hours; primary labs use shared workloads.
Is NVMe over Fabrics included?
We reference it comparatively but FC workloads anchor the week.
What skill level?
Comfortable with CLI metrics and change control; not an introduction to storage.
Experience notes
“The tuning memo template is now mandatory before we touch production LUN layouts.”