Скачать презентацию Track 4 How to build trouble -free large Скачать презентацию Track 4 How to build trouble -free large

ed2b11522875f263078be46eddfa072d.ppt

  • Количество слайдов: 36

Track 4: How to build trouble -free large SANs up to thousand(s) of ports Track 4: How to build trouble -free large SANs up to thousand(s) of ports Dragon Slayer Consulting Marc Staimer, President & CDS [email protected] net 26 April 2004

Agenda l SAN Definition 2004 l Current Large SAN Architectures l Issues w/current architectures Agenda l SAN Definition 2004 l Current Large SAN Architectures l Issues w/current architectures l Eliminating SAN Scaling pain l Summary

Dragon Slayer Background l 7 yrs sales mgt l 10 yrs marketing & bus Dragon Slayer Background l 7 yrs sales mgt l 10 yrs marketing & bus dev • • Storage & SANs 6 years consulting l Launched or participated • 20 products l Paid Consulting • > 70 vendors l Unpaid Consulting • > 200 end users l Known Industry Expert • Speak ~ 5 events/yr • Write ~ 3 trade articles/yr

SAN Definition 2004 SAN Definition 2004

Audience Response Raise your hand if you now have or plan to have within Audience Response Raise your hand if you now have or plan to have within 12 months an all-encompassing SAN infrastructure into the thousands of ports.

Large SAN Architectures l Traditional (a. k. a. Victorian) l Planned/Gated Communities l Urban Large SAN Architectures l Traditional (a. k. a. Victorian) l Planned/Gated Communities l Urban Sprawl

Audience Response By a show of hands, what SAN architecture have you implemented? 1. Audience Response By a show of hands, what SAN architecture have you implemented? 1. Core-to-edge 2. Mesh 3. SAN Islands 4. Not sure

Traditional: a. k. a. Victorian l Mesh • Switch-switch interconnect l Core-to-edge • Guaranteed Traditional: a. k. a. Victorian l Mesh • Switch-switch interconnect l Core-to-edge • Guaranteed hop count & latency l Dual fabric typical for both

Issues with Traditional Approaches l Change management l Guaranteed bandwidth l Fabric disruption propagation Issues with Traditional Approaches l Change management l Guaranteed bandwidth l Fabric disruption propagation

Change Management l Change “No” management • Lot of coordination § Servers, storage, SAN, Change Management l Change “No” management • Lot of coordination § Servers, storage, SAN, cables & facilities • Re-architecting § Switch ports have to be reallocated for ISLs § Zones, cabling, and LUN masking must be redone • Followed up with shakedown & troubleshooting § Sometimes requiring back out of the change

Guaranteed Bandwidth l Lack of user definable Qo. S • Some applications have higher Guaranteed Bandwidth l Lack of user definable Qo. S • Some applications have higher priorities than others

Fabric Disruption Propagation l Fabric disruptions anywhere… • • …propagate throughout the fabric everywhere Fabric Disruption Propagation l Fabric disruptions anywhere… • • …propagate throughout the fabric everywhere RSCNs § Zone changes, add switches or HBAs

Traditional Approaches have led to Urban Sprawl: a. k. a SAN Islands l IT Traditional Approaches have led to Urban Sprawl: a. k. a SAN Islands l IT is dynamic • Most organizations do not plan well l Minimizes disruption effects of change • Doesn’t eliminate disruptions …this This becomes…

Issues with SAN Islands l Limits SAN benefits • • Storage consolidation limited by Issues with SAN Islands l Limits SAN benefits • • Storage consolidation limited by island Management touches expand

Eliminating SAN Scaling Pain: The Market Requirements l Fabric disruptions l Large fabric latency Eliminating SAN Scaling Pain: The Market Requirements l Fabric disruptions l Large fabric latency • Intra-fabric switch ASIC hops l Database bloat l Qo. S l Change management l Correlating storage provisioning, SANs, & policies l Troubleshooting

Fabric Disruptions l RSCNs • Switch, HBA, Zoning l Changes l Market requirement • Fabric Disruptions l RSCNs • Switch, HBA, Zoning l Changes l Market requirement • < fabric disrupts

Intra-Fabric Switch ASIC hops l Hop number affects latency • • Latency is cumulative Intra-Fabric Switch ASIC hops l Hop number affects latency • • Latency is cumulative Affects end user response times l Users demand predictability • Mesh and/or SAN islands = unpredictable § Locality = predictability again • Core-edge = predictable l Market requirement • Minimize latency

SAN Database Bloat l As fabrics get larger • • FSPF databases get larger…and SAN Database Bloat l As fabrics get larger • • FSPF databases get larger…and slower Name services get larger…and slower l Market requirement • Keep databases small

Qo. S l Policy based bandwidth matching • • Providing each application bandwidth based Qo. S l Policy based bandwidth matching • • Providing each application bandwidth based on User defined requirements and thresholds l Market requirement • • Optimize bandwidth Not to waste it

Change Management l Market requirements include • • Automation Negative impact minimization Audit trail Change Management l Market requirements include • • Automation Negative impact minimization Audit trail Change simulation, planning, & validation Correlation of LUN maps, zones, pathing policies Work plans for all of the departments involved Simple, “brain dead” trouble shooting

Correlating Storage Provisioning, SANs, & policies l Efficient storage mgt = < SAN l Correlating Storage Provisioning, SANs, & policies l Efficient storage mgt = < SAN l Market requirements include • • • One interface for both storage &and SAN mgt Policy based Enforcement capable

Troubleshooting l Market requirements include • • • Make it brain-dead simple Make it Troubleshooting l Market requirements include • • • Make it brain-dead simple Make it quick Make it easy AND cheap

Audience Response By a show of hands, which is your worst SAN scaling pain? Audience Response By a show of hands, which is your worst SAN scaling pain? 1. Fabric disruptions 2. Large fabric latency 3. Database bloat 4. Qo. S 5. Change management 6. Storage, SANs, policies correlation 7. Troubleshooting

Solutions that Eliminate SAN Scaling Pain l HBA RSCN switch suppression l Automated change Solutions that Eliminate SAN Scaling Pain l HBA RSCN switch suppression l Automated change mgt software l SAN Masking-a. k. a. SAN routing l SAN segmentation • Planned communities l Qo. S l SAM l Troubleshooting tools

HBA RSCN Switch Suppression l Stops unimportant HBA RSCN disruptions • • From disrupting HBA RSCN Switch Suppression l Stops unimportant HBA RSCN disruptions • • From disrupting other HBAs Significantly < zoning requirements l Vendors include • • QLogic Mc. DATA RSCN

Automated Change Management Software l Plan Change l Predicts Impact l Implements Change l Automated Change Management Software l Plan Change l Predicts Impact l Implements Change l Validates Change l Logs Change History l Correlates • Storage & SAN changes • • • LUNs Zones Pathing Policies l Vendors include • Onaro

SAN Masking-a. k. a. SAN Routing l Analogous to LUN masking l Routes specific SAN Masking-a. k. a. SAN Routing l Analogous to LUN masking l Routes specific data • Between SAN islands • Visibility between specific WWNs l Eliminates disruptions • Between SAN islands SAN Fabric B SAN Fabric A AN VS N 1 A VS l Simplifies management • Both ongoing & change mgt. • Heterogeneous SANs • Address translation (domain & WWN) • Eliminates ATL forced fabric merges • Increases availability 2 l Increases SAN scalability • > switches from 239 to 57, 121 SAN Fabric C

SAN Masking continued l Works over FC • • And IP networks i. FCP SAN Masking continued l Works over FC • • And IP networks i. FCP and FCIP l Vendors include • Mc. DATA SAN Fabric B SAN Fabric A § Eclipse/IPS Cisco • N 1 § Multiprotocol Router AN Brocade A VS • 2 § MDS: VSAN Routing VS • Light. Sand § 8100 SAN Fabric C

SAN Segmentation: a. k. a. Planned Communities l Analogous to large storage controller • SAN Segmentation: a. k. a. Planned Communities l Analogous to large storage controller • Start large & subdivide l One physical fabric • Many logical ones l Vendors include • Cisco § MDS: VSANs • Mc. DATA § Dynamic Partitioning • CNT (04)

Quality of Service: Qo. S l SAN throughput allocation l Based on IT priorities Quality of Service: Qo. S l SAN throughput allocation l Based on IT priorities l Policy based Qo. S l Recognizes App performance • • Requirements differ OLTP > than data migration, etc. l Vendors include • SANdial: Shadow 1400 § Inter & intra-switch • Cisco: MDS § Intra-switch OLTP 100 MB/s Migration 25 MB/s Mc. DATA (04) Warehouse 35 MB/s CNT (04) Email 30 MB/s • •

System Area Management: SAM l SRM + SAN mgt • • • Storage Provisioning System Area Management: SAM l SRM + SAN mgt • • • Storage Provisioning Block & File Heterogeneous Policy based mgt Policy enforcement tools One look & feel l App performance mgt l Optimizes ecosystem l Vendors include • EMC • Softek • App. IQ • HP • IBM • Creekpath • VERITAS • Storability • Tek. Tools • CA

Easier Troubleshooting Tools l Simplified • • • Problem isolation l Vendors include • Easier Troubleshooting Tools l Simplified • • • Problem isolation l Vendors include • Problem resolution Performance issues Cisco § SPAN, r. SPAN • SANdial § Network Performance Analyzer

How Big Can SANs Grow? l Switches • Currently up to 256 ports § How Big Can SANs Grow? l Switches • Currently up to 256 ports § Up to 1024 2 H 2004 l Fabrics • Traditional § 239 switches • 239 x 256 = > 61 K ports • Theoretical (new technologies) § 239 switch domains § 239 switches/domain § 256 ports/switch § = > 14 M ports

Conclusion l SAN Scaling today is painful l New generation software & hardware • Conclusion l SAN Scaling today is painful l New generation software & hardware • Provides pain relief l Test & verify

Thank you. Questions? Thank you. Questions?

Mr. Staimer will be available in the Ask-the-Expert booth in the Exhibit Hall: Monday Mr. Staimer will be available in the Ask-the-Expert booth in the Exhibit Hall: Monday 5 -6 PM