Geek Sync I Capacity Planning for Improved Uptime
Download as PPTX, PDF1 like199 views
The document discusses capacity planning for improved uptime, detailing methods and challenges associated with managing capacity constraints that affect systems and networks. Key areas of focus include addressing storage and network limitations and using various data sources and statistical analyses for effective capacity planning. It also outlines remediation strategies, such as anticipatory and reactive approaches, as well as best practices for documentation and data consistency.
Geek Sync I Capacity Planning for Improved Uptime
- 1. Capacity Planning for Improved Uptime April 13th, 2016
- 2. Overview Analysis Methods / Challenges Why Bother? Points of Constraint Remediation 3 1 2 4
- 3. Why Bother? Everything is running great – and nothing ever changes…
- 4. Why Bother? Capacity Constraints Impact • Unplanned Downtime • Internal User Experience • Budget Overruns / Unplanned Expense • Operational Efficiency • Peace of Mind
- 5. Points of Constraint Common Areas of Focus (Network) • WAN Circuits • Access Layer Uplinks • QoS Buckets • VPN Sessions • CPU • Memory
- 6. Points of Constraint Common Areas of Focus (Systems) • Storage (IOPs) • Storage (Capacity) • Storage (Latency) • Storage (everything else) • CPU • Memory
- 7. Capacity Planning Methods Manual Analysis is Hard • Infrastructure Changes • Shifting Demand / Growth Events • Data Overload • Demand Cycles • Time • Blended Workloads
- 8. Capacity Planning Methods Common Data Sources (Network) • SNMP-derived If / CPU / Memory • Netflow / NBAR • IPSLA • Topology Map
- 9. Capacity Planning Methods Common Data Sources (Systems) • SNMP/WMI derived If / CPU / Memory / Storage • vCenter • Storage Arrays • Dependency Map • SSRS
- 10. Capacity Planning Methods Statistical Analysis • Modeling (Current + Future Known Load) • Trending (Future reflects past) • Forecasting (Current + Unknown Future) • Workload Impact (How does workload impact constraints?)
- 11. Capacity Planning Methods Statistical Analysis • Static / Normalized Datasets • ToD / DoW • Linear Regression • Peak Load
- 12. Remediation Strategies Anticipatory • Use Modeling / Forecasting to pre-purchase • Least risk of service interruption • Greatest risk of financial waste • Particularly with resources defined by step- functions • Most time consuming
- 13. Remediation Strategies Reactive • Use current load baseline(s) • May derate constraints / Use thresholds • Greatest risk of service interruption • Least risk of financial waste
- 14. Remediation Tactics Network • Buy More! Buy Bigger! • Implement QoS / Policing • WAN Op is still a thing • CDN / Edge Caching • AWS Direct Connect
- 15. Remediation Tactics Systems • Buy More! Buy Bigger! • Thin Provisioning • Address VM Sprawl • Storage Tiering • Cold / Offline Storage • Hybrid Cloud Adoption – Capex vs. Opex
- 16. Remediation Best Practices • Document Process and Assumptions • Pick Consistent Data • Run Multiple Models (Best/Worst/Avg) • Reconcile Outcomes with Expectation
- 17. Or find an automated system to do the heavy lifting…