Six Steps to Private Cloud
1 like598 views
The 6 steps to building a private cloud are: 1. Standardize and simplify infrastructure by choosing critical vendors to reduce management points. 2. Drive down I/O consolidation to reduce cables and switches through network collapse and native 10Gbps connectivity. 3. Improve storage efficiency through techniques like thin provisioning, deduplication, and thin replication to reduce costs. 4. Automate infrastructure through tools that configure attributes and rapidly provision resources like hypervisors do for operating systems. 5. Implement uniform scale through a single management entity that can scale resources linearly. 6. Orchestrate resources through self-service delivery, policy-driven provisioning, and model-based management to improve innovation.
Six Steps to Private Cloud
- 1. 6 Step’s to Private Cloud Rob Gee Datacentre Architect
- 2. 3 About Logicalis Group Over $1.2 billion in revenues Operations in Europe, USA, Latin America and Asia Pacific Almost 3,000 employees worldwide Over 6,000 corporate and public sector customers Strong global partnerships with technology leaders - Cisco Systems, HP, IBM, NetApp and CA technologies Logicalis is an internationalIT solutionsand managed services providerwith a breadth of knowledgeand expertisein communicationsand collaboration; datacentreand cloud services; and managed services
- 3. 4 Logicalis Australia At a Glance Officesacrosseastern seaboard (Melbourne, Sydney and Brisbane) Circa $80M revenue Strongservicesmix – 49% in FY12 220 staff in Australia and Malaysia (back-office) Strategy based on leading with differentiated solutionsthat addressbusinesschallenges
- 4. Logicalis Partner Capabilities • Cisco Gold Partner – Advanced Data Centre Architecture – Advanced Borderless Architecture – Advanced CollaborationArchitecture – Cloud Builder Partner – ManagedServices Master • NetApp Gold Partner – Advanced SupportPartner – Cloud Builder Partner • EMC • Citrix – Gold Partner – Citrix Cloud Partner • Microsoft & VMware Partner’s
- 5. Improve business processes 1 Lower overall coststructure 2 3 4 5 6 7 8 Improve or simplify IT infrastructure Building a more secure IT environment Business intelligence Energy-efficientIT infrastructure Finding the right talentto meet ICT priorities Harnessing ICT to gain competitive advantage CIO / ICT priorities in 2013?
- 6. What is your perspective of Cloud Computing? “Once upon a time, there were six blind men. The blind men wished to know what an elephantlookedlike. The six blind men all perceived oneaspectof the elephant and were each right in their own way, but none of them knew what the whole elephantreally looked like.” Your definition of something may depend on your perspective… It’s like a rope! It’s like a wall! It’s like a fan! It’s like a snake! It’s like a spear! It’s virtualization! It’s automation! It’s statelessness! It’s on-demand computing! Cloud Computing It’s a self-service portal! Its “Anything as a Service” (XaaS)! “Once upon a time, there were six IT people. The IT people wished to know what cloud computing looked like. The six IT people all perceived one aspect of the cloud and were each right in their own way, but none of them knew what the whole cloud really looked like.” Elephant It’s like tree!
- 7. Infrastructure Challenges Islands of technology: Implemented in a complex and fragmented nature Hardware & Software out of Support Unpredictable & spiraling support costs Capacity breaks everything Productivity/efficiency challenges Complex architecture only meant to be temporary No “Uniform Scale”
- 8. ‘Stand Up Infrastructure Cost Training Professional Services Staff Resource Blades/Servers Network Ports Storage Media Cables Transceivers Management Full technology shift Services protocol change Addition of secondary technology No Concept of ‘Uniform Scale’ Within the component technology and the ‘stack’ Creates Cost Prohibitive ‘Break Points’ (Further Challenged by higher density technologies such as VDI) $ No Uniform Scale
- 9. Options ? • Bolt on Aftermarket Extension’s • MarginalGain (if any) • Increased Managementand Cost • Refresh Infrastructure • Same Service • Same Challenges • Maybe Different Badge • Move to a Cloud Model • Defined SLA’s • Validated Infrastructure • Reduce Cost / Service Improvement • Delivered via “As a Service” models
- 10. 6 Steps to Private Cloud 1# Standardise & Simply 2# IO Consolidation 3# Storage Efficiency 4# Automation 5# Uniform Scale 6# Orchestration & Management
- 11. TapeBackup NAS (FileBased) SAN (Block Based) Blade Arrays Blade Centre Modules Rack Optimised Computing IP Top of Rack FCP Top of Rack LAN/WANHow do I Automatethis? & How much is wasted? Core Network Typical Chassis Based (6509) Top of Rack Switching IP - 1U 48 Ports (3750/4948) FC – 1U 24 Ports SAN Block I/O,Speed High Throughput DB IBM DS/EMC/HDS/HP NAS Oracle DB/File Services iSCSI More Storage lesscost Traditional Data Centre: Waste
- 12. Converged Datacenter Model TapeBackup LAN/WANEnd to End Unified Fabric TotalNetworkCollapse Lose Native FC Network CompletelyCollapse Infrastructure Improve Transactionsper kW Increase I/O Utilisation LessPower LessCooling LessCabling Blade Arrays Blade Arrays
- 14. 1# Standardise and Simplify Critical Vendor Choice – Time to be strategic…. Data Storagefor a Particular Purpose(SAN/NAS)? Multiple Points of Backup and Data Protection Multiple Points of Management & Control Can you Workflow, Automate and Orchestrate so many vendor products?? Reduce the Points of Management… Reduce Cost & Increase Agility
- 15. 2# Drive Down I/O Consolidation Chassis Mgmt Fibre Channel Blade Switch Mgmt Ethernet Blade Switch Mgmt Ethernet Switch Mgmt FC Switch Mgmt Provide Hypervisor Visibility and/or Pass Through • Reduce Cables • Native 10Gbps – Not for Bandwidth, I/O Consolidation and Simplification (10 Cables to 1) • Cloud Style InfrastructureShould ProvideALL Protocols • Reduce Modules/Ancillary Components • Reduce Switchports Reduce Cost & Increase Utilisation
- 16. Virtual Clones Saveup to 80% using writable virtual copies; only changed data uses capacity Thin Replication Saveup to 95% capacity with our disk-to-disk data protection Snapshot Copies Save80% capacity; only changed data uses storage Storage Tiering Better useof EFD, SAS & SATA disks Deduplication Reclaim up to 95% capacity for full backups; 45% on average Thin Provisioning Liberate 20% to 33% of your capacity 3# Storage Efficiency Reduce Cost & Increase Efficiency
- 17. 4# Automation Does to Hardware as Hypervisors do to Operating Systems Infrastructure Re-Purpose Daily/Nightly Tasks All Compute Attributes Configurable from a Single GUI Openly Configurable via API Rapid Scale, Instant Provision Cloud/IaaS Ready Rapid Provisioning No Assembly Required!! SAN LAN Chassis-9/Blade-2 Server Name:LS-A UUID: 56 4d cd 3f59 5b 61… MAC : 08:00:69:02:01:FC WWN: 5080020000075740 Boot Order:SAN,LAN Chassis-1/Blade-5 Reduce Cost & Decrease Time to Market
- 18. 5# Uniform Scale 1 Management Entity 1 IP Address LAN SAN/NAS ScaleOutSATA RackLikeScale BladeFormFactor Reduce Cost & Provide Linear Scalability
- 19. ‘Stand Up Infrastructure Cost Training Professional Services Staff Resource Blades/Servers Network Ports Storage Media Cables Transceivers Management Full technology shift Services protocol change Addition of secondary technology No Concept of ‘Uniform Scale’ Within the component technology and the ‘stack’ Creates Cost Prohibitive ‘Break Points’ (Further Challenged by higher density technologies such as VDI) $ No Uniform Scale
- 20. Uniform Scale $
- 21. 6# Orchestration & Management Self-service Delivery Policy Driven Provisioning Cisco UCS Director Unified Service Delivery Secure Cloud Container VMsvComputevNetwork Stora ge Tenant B Tenant C Tenant A Model-based Management Virtual Platform Storage Network Compute Spend more time innovating
- 22. Virtual Machines& Compute Storage Networking A Simple Design That Scales At Multiple Layers
- 23. TapeBackup NAS (FileBased) SAN (Block Based) Blade Arrays Blade Centre Modules Rack Optimised Computing IP Top of Rack FCP Top of Rack LAN/WANHow do I Automatethis? & How much is wasted? Core Network Typical Chassis Based (6509) Top of Rack Switching IP - 1U 48 Ports (3750/4948) FC – 1U 24 Ports SAN Block I/O,Speed High Throughput DB IBM DS/EMC/HDS/HP NAS Oracle DB/File Services iSCSI More Storage lesscost Traditional Data Centre: Waste
- 24. Converged Datacenter Model TapeBackup LAN/WANEnd to End Unified Fabric TotalNetworkCollapse Lose Native FC Network CompletelyCollapse Infrastructure Improve Transactionsper kW Increase I/O Utilisation LessPower LessCooling LessCabling Blade Arrays Blade Arrays
- 25. Virtualised at 5 Layers Operating SystemVirtualisation For consolidation, workloadmanagement business continuity and Disaster Recovery Compute Hardware Virtualisation For infrastructurere-purpose andrapidprovision Storage Virtualisation Allows multiple protocol support andhighefficiencies I/O Virtualisation / Unified Fabric Any protocol > Any wire Data Centre Virtualisation 1 Virtual DataCentre physically splitover multiple sites
- 26. Cisco Validated Design’s (CVD’s) Secure Multi-Tenancy • FlexPod (NetApp) • VSPEX (EMC) View Microsoft Exchange Microsoft SharePoint Microsoft SQL Server • Step-by-step instructions for repeatable high-quality deployments • 30+ Validated Designs • Co-operative Support Model • Logicalis Cloud (LVPDC)
- 27. Converged Datacenter Marketplace Hypervisor FlexPod Vblock VSPEX CloudMatrix vStart ComputeNetworkStorageMarket PureFlex Market Leaders Others 2,100 customers since launchin December 2009 (initially known as SMT) 1,000 systems since launchin November2009 Launched in January 2011 Launched in April 2011 Launched in April 2012 Launched in April 2012
- 28. Questions