Virtualization: Doing it right the first time to avoid costly ...

  • 114 views
Uploaded on

 

  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Be the first to comment
    Be the first to like this
No Downloads

Views

Total Views
114
On Slideshare
0
From Embeds
0
Number of Embeds
0

Actions

Shares
Downloads
0
Comments
0
Likes
0

Embeds 0

No embeds

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
    No notes for slide
  • LOOK FOR UPDATED SLIDE WITHOUT the SPECIAL price. Gold & Platinum support options are available!
  • Gold & Platinum support options are available!
  • Shared Storage Pools - As you know, Virtual Machines are comprised of files and in order to leverage the dynamic management features in VMware like HA, VMotion, etc these files must reside on a shared storage architecture. Shared Storage Pools are where one stores the majority of their Virtual Machines and have management implications.
  • FC – PROS: Communicates over a dedicated path and are not at risk of being overloaded. CONS: Cost FCoE – PROS: Unified Fabric saves money in port and cable reduction, and operational simplicity. Great for 10GB environments with existing FC storage. CONS: Cost of new Networking infrastructure. iSCSI – PROS: Great for first shared array, cost and simplicity dominate. Entry-level VMware ESX implementations. Requires less hardware, it uses lower-cost hardware, and more IT staff members are familiar with the technology. CONS: When an iSCSI path is overloaded, the TCP/IP protocol drops packets and requires them to be resent. When a network path carrying iSCSI storage traffic is oversubscribed, a bad situation quickly grows worse and performance further degrades as dropped packets must be resent. NAS (NFS) – PROS: Relies on inexpensive TCP/IP and Ethernet cabling. CONS: Not good for applications over 100MBs 1) Does your application require 100 MBs or less? If so, then iSCSI or NFS over 1 GbE will work well. 2) If your application is more demanding then you will need more bandwidth which means you must deploy of FC.
  • So how much difference does it make in a real example. Let’s compare an EMC CX4-120 to a SATABeast. Over the course of three years the CX4 will consume 48,467 kWs of energy as compared to 28,338 kWs for SATABeast. If you look at the difference between EMC’s number and ours and then add the cascade effect, there is an additional benefit of 30,088 kWs watts saved. Adding both the CAPEX and OPEX numbers together, that puts EMC’s cost at $200,510 versus $76,635 for Nexsan. Nexsan is 62% less expensive. By the way, this is just one 4U in a rack, and it still save the equivalent of the amount of energy that would be used by 2.51 single family homes. Imagine what it looks like when you compare configurations that are larger.
  • DCiE = 47% PUE = 2.13 This was run in Rev 2 of Efficiency TradeOff Tool (not yet released as of April 13, 2009) Assumptions: 1MW data center capacity 50% loaded Traditional UPS Single path power Chilled water Cooling tower Perimeter cooling Non-redundant cooling system No economizer Standby generator
  • Before Here you have a “normal” legacy type data center. The UPS is about 2/3 loaded and the cooling is about 50% loaded. This data center has a fairly low DCiE of about 29%. DCiE (Data center infrastructure efficiency) expresses electrical efficiency of a data center as the fraction of the total electrical power supplied to the data center that is ultimately delivered to the IT load. Expressed this way, efficiency falls between zero and one, and is commonly expressed as percent (100% being perfect efficiency, meaning all power into the data center is delivered to the IT loads). Power to IT/Power In After an IT virtualization project, the scaled down quantity of servers are spread out across the IT racks. The servers can not all be put into a single rack because the room cooling is too far away from the centralized high density. You can see now that the cooling is now running at about ¼ of its capacity while the UPS is running at about one third. These low level puts the power and cooing into very inefficent states of operation as the efficiency lowers as the load lowers generally. The resulting DCiE is now at 16%. PLUS right-sized power and cooling If the project included APC InfraStruXure purchase at the same time as the IT virtualization project they could have right-sized the power and cooling resulting in a much higher 63% DCiE. Another benefit being a high density deployment that take advantage of “close coupled” high efficiency cooling and the ability eliminate 3 IT racks - real estate savings

Transcript

  • 1. Virtualization: Doing it right the first time to avoid costly repeat attempts. Software Licensing, Storage Options, Power and Cooling and Management
  • 2. vSphere 4 Product Licensing Essentials includes support and Essentials Plus requires support.
  • 3. vSphere 4 Product Licensing Support is required and sold separately. vSphere 4 Acceleration Kits/Promotions
  • 4. Microsoft OS Licensing
    • Operating Systems will need to be licensed
    • Microsoft O/S
    Unlimited VM – licensed per CPU Windows Server Datacenter 4 VM per license Windows Server Enterprise 1 VM per license Windows Server Standard
  • 5. Storage
    • Shared Storage Pools
    • Centralized Management
    • Cost and Simplicity
  • 6. Storage Protocols
    • FC (Fibre Channel)
    • FCoE (Fibre Channel over Ethernet)
    • iSCSI
    • NAS (NFS)
    • The answer is based on your environment and or application requirements.
    • http:// www.vmware.com/resources/techresources /
    What will it be with your VMware environment?
  • 7. VMware Useful Resources
    • Product Information - http:// www.vmware.com /products
    • Events - http:// www.vmware.com /events
    • White Papers - http:// www.vmware.com/solutions/whitepapers.html
    • Compatibility Guides - http:// www.vmware.com/resources/guides.html
    • TCO Calculator - http:// www.vmware.com /calculator
    • Success Stories - http:// www.vmware.com /customers/
    • SMB Portal- http:// www.vmware.com/smb
    • Product Downloads - http:// www.vmware.com /download/
    • Solution Tracks Online - http:// www.solutiontracks.com
  • 8. Zophar Sante Nexsan VP Product Marketing Mike Cirrincione Nexsan Sales Manager
      • MicroAge Presents
      • Nexsan
      • Storage Strategies for
      • VMware
  • 9. Agenda for Today
    • New Storage Technology
    • For Aligning VMware
    • Server Cost Reduction Strategies
    • with
    • Storage Cost Reduction Strategies
    • Reducing Storage Power
    • Reducing Storage Space
    • Reducing Storage Hardware
    • Reducing Storage Management
    • Storage Architectures for VMware
    • Single ESX Server
    • Multiple ESX Servers / VMotion
    • Multiple ESX Servers / VMotion / Disaster Recovery
  • 10. New Technology to Reduce Storage POWER
    • Old MAID
    • Reduces application performance for energy savings
    • “ On-or-off” approach to energy savings
    • Usually only supported on SATA drives
    • Minutes to spin up components to respond to IO requests
    • Slow response confines use to just a few applications
    • AutoMAID
    • Benefits of MAID without the limitations!
    • Energy savings with application performance “Green with Speed”
    • Supported on SAS and SATA drives
    • Three levels of energy savings
      • Level 1, up to 20%
      • Level 2, up to 40% (SATA only)
      • Level 3, up to 60%
    • Fast response to 1 st I/O requests
      • Level 1, sub-second
      • Level 2, less than 15 seconds
      • Level 3, less than 30 seconds
    • Fast and flexible response broadens use to include a wide range of applications
  • 11.
    • Using LEGACY technology compared to NEW technology
        • $.15 Kilowatt hour cost
        • 3 year period
        • 84TB
    • Results - 60% less expense (includes CAPEX , OPEX)
    The results when you use New Technology to Reduce Storage POWER LEGACY NEW T3 Year Power kWhs $17,382 $6996
  • 12. New Technology to Reduce Storage SPACE Drives are installed back to back to reduce vibration Drive bottom Drive bottom
      • Density
      • Single 4U chassis
      • Up to 84TB
      • 21TB per U
      • SAS and SATA
      • 4 FC ports
      • 4 iSCSI ports
      • Performance
      • 4 RAID Engines per chassis
      • 4 GB cache per chassis
      • 1100MBs and 40,000 IOPs
    Cooling channel reduce drive heat. New 4X Enterprise RAID Engines Per Chassis
  • 13. The results when you use New technology to Reduce Storage SPACE
    • Using LEGACY technology compared to NEW technology
        • $100 per. U per. month
        • 3 year period
        • FC Ports cost $300 each
        • 84TB SATA (25TB of SAS)
    • Results - 62% less expense (OPEX) – 85% less expensive (Infrastructure CAPEX)
    42TB in 4 U 42TB in 9 U LEGACY NEW TBs per U 4 10.5 U per TB .25 .095 Space Cost per TB x 3yr $75 $28.5 Port Cost 12 x $300 = $3600 4 x $300 = $1200
  • 14. New Technology to Reduce Hardware Components
    • The Results when you use New Technology to Reduce Hardware Components
    • 65% fewer components
    • Less space
    • Less power
    • Fewer points of failure
    • Few things to worry about
    42TB Components LEGACY NEW Power Supplies 6 2 Cooling Fans 18 7 Controllers 6 2 IO Ports 12 4 Enclosures 3 1 Total Hardware 45 16 RACK U 9 4
  • 15. New Technology to Reduce Storage Management
    • The Results when you use New Technology to Reduce Management
    • 1 thing to manage NOT 3 things to manage
    • Less effort to scale
    • Less to worry about
    LEGACY NEW
  • 16. Single ESX Server ESX Virtual Server Host Guests New Storage Technology for A Single ESX Server
    • SAS (Speed) and SATA (Capacity)
    • 3TB to 200TB
    • Dual / Redundant FC / iSCSI connections
    • Active / Active Data Paths
      • Host Port Redundancy
      • Network connection Redundancy
      • Storage controller Redundancy
    • 3TB to 200TB
  • 17. New Storage Technology for Single Site - Multiple ESX Servers with VMotion VMotion ESX Virtual Server Host Guests Guests ESX Virtual Server Host
    • Server Redundancy
    • SAS (Speed) and SATA (Capacity)
    • 3TB to 200TB
    • Dual / Redundant FC / iSCSI connections
    • Active / Active Data Paths
      • Host Port Redundancy
      • Network connection Redundancy
      • Storage controller Redundancy
    • 3TB to 200TB
  • 18. New Storage Technology for Multi-Site ESX Servers / VMotion / Disaster Recovery VMotion ESX Virtual Server Host Guests Guests ESX Virtual Server Host
    • Server Redundancy, Data Redundancy, Storage Redundancy, Site Redundancy
    • 5 Nines = 99.999 HA
    • SAS (Speed) and SATA (Capacity)
    • 3TB to 1000TB
    Site A Site B
  • 19.
    • 8 white papers
    • 5 - 10 minute briefing papers
    • Value Calculator
    Here is some additional Nexsan information available from MicroAge
  • 20. Conclusion
    • New Storage Technology
    • For Aligning VMware
    • Server Cost Reduction Strategies
    • with
    • Storage Cost Reduction Strategies
    • Reducing Storage Power – 60%
    • Reducing Storage Space – 62%
    • Reducing Storage Hardware – 65%
    • Reducing Storage Management – 66%
    • Storage Architectures for VMware
    • Single ESX - YES
    • Multiple ESX / Vmotion YES
    • Multiple ESX / VMotion / Disaster Recovery YES
    • Contact MicroAge to Learn More
  • 21. Nexsan
    • Who are We
    • Nexsan was founded in 1999
    • Over 24,000 storage systems sold into over 60 countries
    • We ship nearly 6000TB per month
    • What do we Do
    • Provide data storage solutions
    • SMB to Large Enterprise
    • All server environments
  • 22. Confidential Here are some Nexsan Customers
  • 23.  
  • 24. Common Mistake Failure to adjust your power and cooling for your new infrastructure may result in decreased efficiency
  • 25. Defining Data Center Infrastructure Efficiency The percent of your input power that gets to the IT loads The rest goes to power, cooling, and lighting equipment DCiE Data center infrastructure efficiency IT load power Total data center input power =
  • 26. POWER system COOLING system The cooling system uses the most energy in the data center Data Center Physical Infrastructure IT
  • 27. Go Go to latest online version of this tool Click to resume presentation mode Done
  • 28. Right-sizing improves power use AND efficiency EFFICIENCY POWER USE Always goes down with virtualization Always goes down, UNLESS power/cooling is downsized to match load AFTER virtualization ROOM cooling Non-scalable UPS Unused rack space Servers Unused POWER capacity Unused COOLING capacity ROOM cooling Non-scalable UPS BEFORE virtualization Power use DOWN Due to consolidated servers 29% Due to losses from excess capacity DCiE LOW 16% Due to losses from GREATER excess capacity DCiE LOWER PLUS right-sized power and cooling Power use down MORE Due to RIGHT-SIZED power/cooling Scalable ROW cooling 63% Due to RIGHT-SIZED power/cooling DCiE HIGHER Scalable UPS