SlideShare ist ein Scribd-Unternehmen logo
1 von 24
Downloaden Sie, um offline zu lesen
COST-EFFECTIVE
SOLUTIONS FOR
BIG DATA WITH HADOOP
Matt Kimball
Product Marketing Manager

Dell World 2012
Austin, Texas
December 13, 2012
WHAT‘S BIG DATA?

                               BIG DATA
Data sets sized beyond the ability of commonly
used tools to capture, manage, and process the
         data within a tolerable time.




                                                                             BIG DATA Technologies
                                                                    A new generation of technologies designed to
                                                                     economically extract value from very large
                                                                         volumes of a wide variety of data.



     Definitions derived from Gartner, IDC, and NIST documents
 2 | Cutting Big Data Down to Size | Dell World – December 2012 |
DROWNING IN DATA?

                                                                            Proliferation of
                                                 Increasing data
                                                                             unstructured
                                                   complexity
                                                                                  data




                         Growing                                                             Surge in
                        numbers of                                                         external data
                         volumes                                        2.5                  streams
                                                                    quintillion
                                                                     bytes of
                                                                   data created
                                                                      daily*


3 | Cutting Big Data Down to Size | Dell World – December 2012 |
WHAT‘S HADOOP?

                Framework for distributed processing of large data sets across
                  clusters of computers using a simple programming model
       • Scalable, fault-tolerant, distributed system for data storage and processing
       • Foundation of rapidly growing ecosystem of supporting projects
       • Created by developers to handle intractable problem (Cloudera, Ebay, Facebook,
         Hortonworks, Linkedin, Twitter, and Yahoo!)
       • Open source project hosted under the Apache Software Foundation




    http://hadoop.apache.org/

4 | Cutting Big Data Down to Size | Dell World – December 2012 |
HOW IS HADOOP BEING USED?


                              INDUSTRY TERM                           VERTICAL       INDUSTRY TERM

                             Social network analysis                    Web       Clickstream sessionization




                                                                                                               PROCESSING
                                                                       Media
   ANALYTICS
   ADVANCED




                              Content optimization                                Clickstream sessionization




                                                                                                                  DATA
                                 Network analytics                     Telco              Mediation

                      Loyalty and promotions analysis                  Retail            Data factory

                                   Fraud analysis                     Financial      Trade reconciliation

                                   Entity analysis                     Federal             SIGINT




Source: Cloudera

   5 | Cutting Big Data Down to Size | Dell World – December 2012 |
WHO IS USING, CONTRIBUTING, AND BUILDING FOR HADOOP?

10Gen                                          Concurrent            Klout                SnapLogic
Accela                                         Datameer              LexisNexis           StackIQ
Adobe                                          eBay                  LinkedIn             StumbleUpon
Alibaba.com                                    eHarmony              Microsoft            SunGard
Amazon.com                                     EMC                   NetApp               Talend
AOL                                            EnterpriseDB          Netflix              Telenav
AT&T                                           Google                Nokia                Teradata
Baidu                                          Groupon               NTT Communications   The New York Times
Bank of America                                Hortonworks           Orbitz               Trend Micro
bitly                                          Hulu                  Pandora              Twitter
Booz Allen Hamilton                            IBM                   Pervasive            Tynt
CBS                                            IIIT Hyderabad        Rackspace Hosting    Vertica
CBS Interactive                                Intel                 Samsung              Visa
Cisco                                          Jaspersoft            SGI                  VoltDB
CNET                                           JPMorgan Chase        Sky                  Yahoo




Source: Hortonworks
                                                         And of course AMD!
http://wiki.apache.org/hadoop/PoweredBy

6 | Cutting Big Data Down to Size | Dell World – December 2012 |
WHY WAS HADOOP CREATED?

                             Exploding data volumes                                       Dramatic changes in enterprise data
                                                                               LEADS TO
                                          and types                                       management

                    DIGITAL
                   CONTENT
                                                                     OPERATIONAL                               •   Extract more value from
                                                                                                 NEW
                                                   WEB                  DATA                                       more data more cost
                                                                                             OPPORTUNITY
                                                   LOGS                                        FOR USING           effectively with greater
                   SOCIAL
     FILES         MEDIA                                       SMART                         CREATED DATA          flexibility
                                                               GRIDS
                                                                                                               •   Deep analysis
                                    TRANSACTIONAL                                          HARD PROBLEMS       •   Exhaustive and detailed
                                        DATA                                                                   •   Sophisticated algorithms
          AD                                                                                                   •   Quick results
      IMPRESSIONS                                                   R&D
                                                                    DATA                                       •   Any type any kind
                                                                                              BIG DATA         •   From any source
                                                                                                               •   Structured and unstructured
  • It’s difficult to handle data this diverse at this scale                                                   •   At scale
  • Traditional platforms can’t keep pace


Source: Cloudera

 7 | Cutting Big Data Down to Size | Dell World – December 2012 |
TRENDS AROUND BIG DATA
                                          Enterprise and Government organizations faced with
                                             analyzing large amounts of unstructured data.


Big Data Trends:
• Low acquisition costs
• Servers that can effectively parallelize workloads
• Cost effectively scale out Hadoop systems to
  meet expanding data processing and analysis
  needs
• Compute density to handle large amounts of data




                            Data Driven Business Decisions and Actions

 8 | Cutting Big Data Down to Size | Dell World – December 2012 |
AMD AND DELL FOR YOUR HADOOP CLUSTERS


             CPU                      • Medium clock speeds
                                                                     Core Density for
                                                                       Parallelism
   Form Factor                        • Two sockets
                                                                         Three P
                                                                       Equilibrium
          Power                       • Concern as cluster grows
                                                                        Scale Out
                                                                        Capability
            RAM                       • 48GB to keep CPUs busy
                                                                     Manageability
        Storage                       • 1 spindle per core – SATA
                                                                      Cost Per Node
       Network                        • 1GB growing to InfiniBand™



10 | Cutting Big Data Down to Size | Dell World – December 2012 |
AMD OPTERON™ SERIES PROCESSORS FOR HADOOP
    LOWEST TCO FOR DELL CUSTOMERS

                                CapEx
Hadoop Deployment Scenario




                                OpEx                    TCO
        24 TB Data

        I/O Bound
         10 Users




                             Core Count
                                                                       C6105
                                                                     AMD 4274HE
                              Scale Out                 Perf          64GB RAM




 12 | Cutting Big Data Down to Size | Dell World – December 2012 |
AMD HADOOP SCENARIO TCO
        $800,000
                               $605,120                                                                $705,040
                       $401,500                                                                                                              Up to 34% Lower
        $600,000                                                            CapEx                                                            Hardware Costs11
                                                                                                  $422,842
                                                                                                                        Dell C6105
        $400,000                                                                                                        HP SL230s
                                                                                                                        Gen8                 Up to 40% Lower*
        $200,000
                                                                     $87,920                                                                           CapEx
                                               $0                                    $12,000
                                                           $18,342              $3,000
                $0
                       Server HW cost     Server SW cost   Rack switch cost Rack + install cost            CapEx



                                                                                                       $183,977
            $200,000
                                                                            OpEx                $142,267
                                                                 $105,195
                                                                                                                                               Up to 23% Lower
            $150,000
            $100,000
                                     $24,479
                                                                                                                       Dell C6105                       OpEx*
                                                    $19,583                           $34,720                          HP DL360e
              $50,000         $15,773
                                               $12,619                         $8,680                                  Gen8
                     $0
                           Server power      Overhead      Mgmt OH/year        Footprint cost        OpEx/yr
                             cost/year     power cost/year

                                                    Up to 32% lower 3 year TCO -                           $407,329 savings*
                                                                                    * See slide titled ” Substantiation | Hadoop TCO Calculations” for assumptions and calculations

13 | Cutting Big Data Down to Size | Dell World – December 2012 |
HADOOP CONFIGURATION COMPARISON
DELL POWEREDGE C6105 VS HP DL360E GEN8




           Dell™ PowerEdge™ C6105 (4 nodes)                                        HP DL360e Gen8
             CPU / Node                  AMD Opteron™ 4274HE (2)                       CPU       Intel Xeon E5-2440 (2)

            Memory / Node                64 GB (8x8 1333MHz)                           Memory    64 GB (8x8 1333MHz)

            Storage / Node               3x 2TB SATA                                   Storage   3 x 2TB LFF SATA

            Network / Node               1GBe                                          Network   1GBe

            PSU / Node                   1100W                                         PSU       750W


                                                                  80 Servers
                                               1 42U rack (C6105) vs. 4 42U racks (DL360e Gen8)
                                       Mellanox SX2016 10GBe switches: 2 for C6105 vs. 8 for DL360e Gen8

                                                                    Hadoop Projects Software




14 | Cutting Big Data Down to Size | Dell World – December 2012 |
EUROPEAN AEROSPACE PROJECT HADOOP SYSTEM

                                   Astronomy Mission Mapping Our Galaxy

    •      4 service nodes
    •      Cluster infrastructure – 5 Nodes
    •      94 x C6105 – 742 Sockets
    •      4x MB w/ 2x 4274HE, 24GB, 3x 3TB
    •      Dell|Force10 + Dell switches




                                                                    2013 Production



15 | Cutting Big Data Down to Size | Dell World – December 2012 |
DELL CLOUDERA SOLUTIONS


          The Dell | Hadoop Big Data Solution
          •     Cloudera’s Distribution for Hadoop
          •     Dell PowerEdge™ C hardware, including the C6105 and C6145
                http://dell.cloudera.com/




      http://www.mellanox.com/webinars/Accelerating-Big-Data-Over-Hadoop-With-InfiniBand/


16 | Cutting Big Data Down to Size | Dell World – December 2012 |
THANK YOU



 For more information on AMD Big Data
 solutions, please visit;
               www.AMD.com/BigData




17 | Cutting Big Data Down to Size | Dell World – December 2012 |
DISCLAIMER
  The information presented in this document is for informational purposes only and may contain technical inaccuracies, omissions and
  typographical errors.

  The information contained herein is subject to change and may be rendered inaccurate for many reasons, including but not limited to product
  and roadmap changes, component and motherboard version changes, new model and/or product releases, product differences between
  differing manufacturers, software changes, BIOS flashes, firmware upgrades, or the like. AMD assumes no obligation to update or otherwise
  correct or revise this information. However, AMD reserves the right to revise this information and to make changes from time to time to the
  content hereof without obligation of AMD to notify any person of such revisions or changes.

  AMD MAKES NO REPRESENTATIONS OR WARRANTIES WITH RESPECT TO THE CONTENTS HEREOF AND ASSUMES NO RESPONSIBILITY FOR ANY
  INACCURACIES, ERRORS OR OMISSIONS THAT MAY APPEAR IN THIS INFORMATION.

  AMD SPECIFICALLY DISCLAIMS ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR ANY PARTICULAR PURPOSE. IN NO EVENT
  WILL AMD BE LIABLE TO ANY PERSON FOR ANY DIRECT, INDIRECT, SPECIAL OR OTHER CONSEQUENTIAL DAMAGES ARISING FROM THE USE OF
  ANY INFORMATION CONTAINED HEREIN, EVEN IF AMD IS EXPRESSLY ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

  ATTRIBUTION
  © 2012 Advanced Micro Devices, Inc. All rights reserved. AMD, the AMD Arrow logo, Radeon, and combinations thereof are
  trademarks of Advanced Micro Devices, Inc. Other names and logos are used for informational purposes only and may be
  trademarks of their respective owners.




18 | Cutting Big Data Down to Size | Dell World – December 2012 |
BACK-UP




19 | Cutting Big Data Down to Size | Dell World – December 2012 |
REFERENCE CONFIGURATION




            Server: Dell™ PowerEdge™ C6105 (4-nodes)                  Software:
            CPU/node:                       AMD Opteron™ 4276HE (4)   OS:             CentOS 5.x (or Red Hat)
            Memory/node:                    64 GB (8x8 1333MHz)       Hadoop
            Storage/node:                   3 x 2TB SATA              Distribution:   Cloudera
            Network/node:                   1 GBe                     JAVA:           Sun Oracle JVM 1.6u2x

            PSU/node:                       1100W                     Monitoring:     Nagios, Ganglia
                                                                      Other:          Hadoop Projects Software


                       Coming soon: AMD Opteron™ 4300 series processor support


20 | Cutting Big Data Down to Size | Dell World – December 2012 |
AMD’S PILOT HADOOP ARCHITECTURE




21 | Cutting Big Data Down to Size | Dell World – December 2012 |
HADOOP TCO CALCULATIONS




22 | Cutting Big Data Down to Size | Dell World – December 2012 |
HADOOP FOR TELCO

                                                                                4 x Motherboard system
                                                                    Hadoop on   Each Motherboard:
                                                                      C6105     • 2 x4274HE Proc.
                                                                                • 4 x 16GB RAM
                                                                                • 4 x 2TB HDD




23 | Cutting Big Data Down to Size | Dell World – December 2012 |
SUBSTANTIATION AND FOOTNOTES

   1.   SVR-217 - AMD Opteron™ 6300 Series processors are expected to have up to 37% higher java performance/watt than AMD Opteron 6200 Series processors.. Estimate based on
         preliminary measurements of java performance/watt in AMD labs as of August 2012 comparing AMD Opteron™ processor Models 6380 and 6278. SVR-217

   2. SVR-168 - AMD Opteron™ 6300 Series processors are expected to have up to 24% higher Java performance than AMD Opteron 6200 Series processors. Estimate based on
       preliminary measurements of server side Java performance in AMD labs as of August 30, 2012. 1,199,838 operations per second using 2 x AMD Opteron™ processors Model
       6278. 1,489,668 operations per second using 2 x AMD Opteron™ processors Model 6380. SVR-168

   3.   SVR-205 - AMD Opteron™ 6300 processors are expected to offer up to 8% better integer performance than AMD Opteron™ 6200 processors. Estimate based on preliminary
        SPECint®_rate2006 results in AMD labs as of August 2012 comparing AMD Opteron™ processor Models 6380 and 6278.

   4.   SVR-216 - AMD Opteron™ 6300 processors are expected to offer up to 7% better floating point performance than AMD Opteron™ 6200 processors. Estimates based on
         preliminary SPECfp®_rate2006 measurements in AMD labs as of August 2012 comparing AMD Opteron™ processor Models 6380 and 6278. SVR-216

   5. SVR-255 - The Dell R815 lowers TCO for virtualization by up to 17% versus the HP DL560 Gen8. TCO calculation based on virtualization scenario with 1600 physical servers
       consolidated down to 40 with each physical server running 40 virtualized machines (12GB RAM each). This is supported by two racks of (20) servers with two Mellanox
       SX1016 10GBe 64-port switches. The cost per rack with installation is $3,000 based on http://www.intel.com/technology/eep/datacenter.pdf as of 9/10/12 and (2) Mellanox
       SX1016 10GBe 64-port switches cost $10,990 each and have a typical power consumption of 113 watts as of 9/10/12 at
       http://www.colfaxdirect.com/store/pc/viewPrd.asp?idproduct=1377. Power costs based on $0.10 per kWhr and using 976 watts per for the Dell R815 server and 588 watts
       for the HP DL560p Gen8 server according to the Dell Energy Smart Solution Advisor (ESSA) tool with a memory-intensive workload set at 100% CPU loading and an input
       voltage of 220 volts as well as the HP Power Advisor Tool set at 100% CPU loading and an input voltage of 220 volts as of 9/24/12 based on the configurations below. The PUE
       is 1.8. Assuming 28 square feet per rack according to http://www.energystar.gov/ia/partners/prod_development/downloads/Additional_FAQs_LightingCooling_Racks.pdf as
       of 9/11/12, and the average fully burdened cost of real estate at $310 per square foot per year based on
       http://www.vmware.com/technology/whyvmware/calculator/index.php as of 9/10/12. Management overhead yearly cost per server based on $81k salary with 1.25
       multiplier for a fully loaded expense (benefits, taxes, etc) of $101,250. $1315 per server derived from dividing the fully loaded expense by 77 servers being managed by IT
       administrator, based on page 4 at http://www.vmware.com/files/pdf/vmware-solution-opex-reducing-opex-wp-en.pdf and http://www.indeed.com/salary/Windows-Server-
       Administrator.html. VMware vSphere license is $3495 per CPU as of 9/25/12 at www.vmware.com. R815 with four AMD Opteron™ processor Model 6274, 512GB RAM, 2 x
       1TB SATA drives, two 10GBe NICs, and 3yr base warranty cost $20140 as of 9/24/12 at www.dell.com. HP DL560p Gen8 with four Intel Xeon processor Model E5-4620, 512GB
       RAM, 2 x 1TB SATA drives, two 10GBe NICs, and 3yr base warranty cost $30948 as of 9/24/12 at www.hp.com. The 3yr TCO using AMD Opteron™-based servers is $1,811,446.
       The 3yr TCO using Intel Xeon-based server is $2,170,350. Assuming customer already has agreement for client license and run time license, so not including Windows Server
       license. Actual results will vary. SVR-255




24 | Cutting Big Data Down to Size | Dell World – December 2012 |
SUBSTANTIATION AND FOOTNOTES, CONTINUED
 6. SVR-254 - The Dell R815 offers up to a 35% lower server hardware acquisition cost than the HP DL560 Gen8 for virtualization configurations. R815 with four AMD Opteron™ processor Model 6274,
      512GB RAM, 2 x 1TB SATA drives, two 10GBe NICs, and 3yr base warranty cost $20140 as of 9/24/12 at www.dell.com. HP DL560 Gen8 with four Intel Xeon processor Model E5-4620, 512GB RAM, 2 x
      1TB SATA drives, two 10GBe NICs, and 3yr base warranty cost $30948 as of 9/24/12 at www.hp.com. SVR-254

 7. SVR-213 - Virtualized AMD-based servers deliver up to 12% higher performance and up to 23% better price performance than the competition. Testing with VMware ESX in today’s typical virtualized
      environments consisting of mid-range processor and memory configurations, reveals that AMD Opteron processors lead in both price as well as price/performance versus Intel-based systems.
      Virtualized AMD-based servers deliver up to 12% higher performance and up to 23% better price performance than the competition. Test run VMware ESX with DVD Store, an open source
      performance tool that includes web and database servers to emulate an ecommerce environment.. Servers running increasing number of VMs from 2 to 32 with a fixed, utilization rate of ~25% .
      Systems include: 2X AMD Opteron™ processors Model 6274 in HP DL385 Gen8 server running 4,233 Operations per Minute (OPM) for 2 VMs to 1,242 OPM for 32 VMs with 96GB (12x8 GB DDR3-
      1333) memory, 1x15K 500GB SAS HD, base warranty and costing $7076 for the hardware and $6990 for VMware vSphere Enterprise Plus software; 2 x Intel Xeon E5-2665 processors in a HP DL380p
      Gen 8 server running 3,769 OPM for 2 VMs to 1,178 OPM for 32 VMs, 96GB (12x8 GB DDR3-1333) memory, 1x15K 500GB SAS HD, base warranty and costing $9,338 for hardware and $6990 for
      VMware vSphere Enterprise Plus software . System prices as of 7/5/2012 http://www.hp.com and software as of 7/5/2012 http://vmware.com. SVR-213

 8. SVR-303 - The Dell C6145 lowers TCO for VDI by up to 28% versus the HP SL230s Gen8. For the Dell C6145, the TCO calculation is based on a VDI scenario with 8000 users, 200 VDI sessions per server,
      and 2.5GB physical RAM per session, which is supported by one rack of (20) two-node 4P servers and two Mellanox SX1016 10GBe 64-port switches. For the HP SL230s Gen8, the TCO calculation is
      based on a VDI scenario with 8000 users, 100 VDI sessions per server, and 2.5GB physical RAM per session, which is supported by one rack of (80) 2P servers and two Mellanox SX1016 10GBe 64-
      port switches. The cost per rack with installation is $3,000 based on http://www.intel.com/technology/eep/datacenter.pdf as of 9/10/12 and (2) Mellanox SX1016 10GBe 64-port switches cost
      $10,990 each and have a typical power consumption of 113 watts as of 9/10/12 at http://www.colfaxdirect.com/store/pc/viewPrd.asp?idproduct=1377. Power costs based on $0.10 per kWhr and
      using 1650 watts per for the Dell C6145 server and 417 watts for the HP SL230s server plus 182 watts for the C6500 enclosure according to the Dell Energy Smart Solution Advisor (ESSA) tool with a
      memory-intensive workload set at 100% CPU loading and an input voltage of 220 volts as well as the HP Power Advisor Tool set at 100% CPU loading and an input voltage of 220 volts as of 9/26/12
      based on the configurations below. The PUE is 1.8. Assuming 28 square feet per rack according to
      http://www.energystar.gov/ia/partners/prod_development/downloads/Additional_FAQs_LightingCooling_Racks.pdf as of 9/11/12, and the average fully burdened cost of real estate at $310 per
      square foot per year based on http://www.vmware.com/technology/whyvmware/calculator/index.php as of 9/10/12. Management overhead yearly cost per server based on $81k salary with 1.25
      multiplier for a fully loaded expense (benefits, taxes, etc) of $101,250. $1315 per server derived from dividing the fully loaded expense by 77 servers being managed by IT administrator, based on
      page 4 at http://www.vmware.com/files/pdf/vmware-solution-opex-reducing-opex-wp-en.pdf and http://www.indeed.com/salary/Windows-Server-Administrator.html. XenDesktop VDI software
      costs $134.75 per VDI session as of 9/10/12 provided by Sales at Insight. Dell C6145 with two server nodes of four AMD Opteron™ processor Model 6274, 512GB (32x16 1600MHz) RAM, 2 x 1TB
      SATA drives, one 10GBe NICs, and 3yr 24x7 warranty cost $43978 as of 10/4/12 according to Dell Sales. HP SL230s Gen8 with two Intel Xeon processor Model E5-2665, 256GB (16x16 1600MHz)
      RAM, 2 x 1TB SATA drives, one 10GBe NICs, and 3yr 24x7 warranty cost $19669 plus $2008 for the HP C6500 enclosure as of 10/4/12 at www.hp.com. The 3yr TCO using AMD Opteron™-based
      servers is $2,323,545. The 3yr TCO using Intel Xeon-based servers is $3,205,690. Assuming customer already has agreement for client license and run time license, so not including Windows Server
      license. Actual results will vary. SVR-303




25 | Cutting Big Data Down to Size | Dell World – December 2012 |
SUBSTANTIATION AND FOOTNOTES, CONTINUED
9. SVR-302 - The Dell C6145 offers up to a 44% lower server hardware acquisition cost than the HP SL230s Gen8 for VDI configurations. Dell C6145 with two server nodes of four
    AMD Opteron™ processor Model 6274, 512GB (32x16 1600MHz) RAM, 2 x 1TB SATA drives, one 10GBe NICs, and 3yr 24x7 warranty cost $43978 as of 10/4/12 according to
    Dell Sales. HP SL230s Gen8 with two Intel Xeon processor Model E5-2665, 256GB (16x16 1600MHz) RAM, 2 x 1TB SATA drives, one 10GBe NICs, and 3yr 24x7 warranty cost
    $19669 as of 10/4/12 at www.hp.com. SVR-302

10. SVR-301 - The Dell C6105 lowers TCO for Hadoop by 32% versus the HP DL360e Gen8. TCO calculation based on Hadoop scenario with 24TB dataset being analyzed by 10
     users and being processed by an I/O bound application. This is supported by one rack of 79 Dell C6105 slave nodes and 1 Dell C6105 master node along with two Dell
     Force10 S60 High-Performance 1/10 GbE switches as well as four racks of 79 HP DL360e Gen8 slave nodes and 1 HP DL360e Gen8 master node along with two Mellanox
     SX1016 10GBe 64-port switches per rack. The cost per rack with installation is $3,000 based on http://www.intel.com/technology/eep/datacenter.pdf as of 9/10/12.
     Mellanox SX1016 10GBe 64-port switches cost $10,990 each and have a typical power consumption of 113 watts as of 9/10/12 at
     http://www.colfaxdirect.com/store/pc/viewPrd.asp?idproduct=1377. Dell Force10 S60 High-Performance 1/10 GbE switches cost $9171 each and have a typical power
     consumption of 156 watts as of 9/26/12 at www.dell.com and http://www.networkcomputing.com/next-gen-network-tech-center/force10s-deep-buffer-switch/229501471.
     Power costs based on $0.10 per kWhr and using 889 watts per for the Dell C6105 with four server nodes in one system and 338 watts for the HP DL360e Gen8 server with a
     transactional workload set at 100% CPU loading according to the Dell Energy Smart Solution Advisor (ESSA) tool with a transactional workload set at 100% CPU loading and
     an input voltage of 220 volts and the HP Power Advisor tool set at 100% CPU loading and an input voltage of 220 volts as of 9/26/12 based on the configurations below. The
     PUE is 1.8. Assuming 28 square feet per rack according to
     http://www.energystar.gov/ia/partners/prod_development/downloads/Additional_FAQs_LightingCooling_Racks.pdf as of 9/11/12, and the average fully burdened cost of
     real estate at $310 per square foot per year based on http://www.vmware.com/technology/whyvmware/calculator/index.php as of 9/10/12. Management overhead yearly
     cost per server based on $81k salary with 1.25 multiplier for a fully loaded expense (benefits, taxes, etc) of $101,250. $1315 per server node derived from dividing the fully
     loaded expense by 77 servers being managed by IT administrator, based on page 4 at http://www.vmware.com/files/pdf/vmware-solution-opex-reducing-opex-wp-en.pdf
     and http://www.indeed.com/salary/Windows-Server-Administrator.html. Open Source software is being used so there is no cost. Each of the four Dell C6105 server nodes
     has two AMD Opteron™ processor Model 4274HE, 64GB RAM, 3 x 2TB SATA drives, 1GBe, and 3yr base warranty with a total system cost of $20075 ($5019 each node) as of
     10/3/12 from Dell Sales. HP DL360e with two Intel Xeon processor Model E5-2440, 64GB RAM, 3 x 2TB SATA drives, 1GBe, and 3yr base warranty cost $7564 as of 9/21/12
     at www.hp.com. The 3yr TCO using AMD Opteron™-based servers is $849,643. The 3yr TCO using Intel Xeon-based servers is $1,256,972. Actual results will vary. SVR-301.

11. SVR-300 - The Dell C6105 offers up to a 34% lower scale-out server hardware acquistion cost than the HP DL360e Gen8 for Hadoop configurations. Each of the four Dell C6105
     server nodes has two AMD Opteron™ processor Model 4274HE, 64GB RAM, 3 x 2TB SATA drives, 1GBe, and 3yr base warranty with a total system cost of $20075 ($5019 each
     node) as of 10/3/12 from Dell Sales. HP DL360e with two Intel Xeon processor Model E5-2440, 64GB RAM, 3 x 2TB SATA drives, 1GBe, and 3yr base warranty cost $7564 as
     of 9/21/12 at www.hp.com. SVR-300




26 | Cutting Big Data Down to Size | Dell World – December 2012 |

Weitere ähnliche Inhalte

Was ist angesagt?

Smarter Computing in a New Era of IT - Dr. Gururaj Rao
Smarter Computing in a New Era of IT - Dr. Gururaj RaoSmarter Computing in a New Era of IT - Dr. Gururaj Rao
Smarter Computing in a New Era of IT - Dr. Gururaj RaoJyothi Satyanathan
 
Tackling big data with hadoop and open source integration
Tackling big data with hadoop and open source integrationTackling big data with hadoop and open source integration
Tackling big data with hadoop and open source integrationDataWorks Summit
 
Unlocking Value in the Fragmented World of Big Data Analytics (POV Paper)
Unlocking Value in the Fragmented World of Big Data Analytics (POV Paper)Unlocking Value in the Fragmented World of Big Data Analytics (POV Paper)
Unlocking Value in the Fragmented World of Big Data Analytics (POV Paper)Cisco Service Provider Mobility
 
Big data datacrunch
Big data datacrunchBig data datacrunch
Big data datacrunchReseau'Nable
 
Dr. Shahbaz Ali, CEO at Tarmin - Business Transformation in the Age of Big Data
Dr. Shahbaz Ali, CEO at Tarmin - Business Transformation in the Age of Big DataDr. Shahbaz Ali, CEO at Tarmin - Business Transformation in the Age of Big Data
Dr. Shahbaz Ali, CEO at Tarmin - Business Transformation in the Age of Big DataGlobal Business Events
 
Quantum Leap in Open Source Collaboration
Quantum Leap in Open Source CollaborationQuantum Leap in Open Source Collaboration
Quantum Leap in Open Source CollaborationHarold Teunissen
 
Towards Expertise Modelling for Routing Data Cleaning Tasks within a Communit...
Towards Expertise Modelling for Routing Data Cleaning Tasks within a Communit...Towards Expertise Modelling for Routing Data Cleaning Tasks within a Communit...
Towards Expertise Modelling for Routing Data Cleaning Tasks within a Communit...Umair ul Hassan
 
BI Forum 2009 - BI Mega Trends
BI Forum 2009 - BI Mega TrendsBI Forum 2009 - BI Mega Trends
BI Forum 2009 - BI Mega TrendsOKsystem
 
Prc open stack conf aug 2012 cox v1
Prc open stack conf aug 2012 cox v1Prc open stack conf aug 2012 cox v1
Prc open stack conf aug 2012 cox v1OpenCity Community
 
Collaboration is Happening
Collaboration is HappeningCollaboration is Happening
Collaboration is HappeningHarold Teunissen
 
Big Data World Forum
Big Data World ForumBig Data World Forum
Big Data World Forumbigdatawf
 
Data centerconsolidation10
Data centerconsolidation10Data centerconsolidation10
Data centerconsolidation10IDGA
 
Self-service Linked Government Data
Self-service Linked Government DataSelf-service Linked Government Data
Self-service Linked Government DataFadi Maali
 
Ib Ms Vision For A Dynamic Infrastructure
Ib Ms Vision For A Dynamic InfrastructureIb Ms Vision For A Dynamic Infrastructure
Ib Ms Vision For A Dynamic Infrastructuresimonarden
 
Empowering the Business with Agile Analytics
Empowering the Business with Agile AnalyticsEmpowering the Business with Agile Analytics
Empowering the Business with Agile AnalyticsInside Analysis
 
Externalization Trend
Externalization TrendExternalization Trend
Externalization TrendNigel Green
 
Enterprise 2.0 Summit 2012 Closing Keynote - Next-Generation Ecosystems And i...
Enterprise 2.0 Summit 2012 Closing Keynote - Next-Generation Ecosystems And i...Enterprise 2.0 Summit 2012 Closing Keynote - Next-Generation Ecosystems And i...
Enterprise 2.0 Summit 2012 Closing Keynote - Next-Generation Ecosystems And i...Dion Hinchcliffe
 
Enterprise Energy Management using a Linked Dataspace for Energy Intelligence
Enterprise Energy Management using a Linked Dataspace for Energy IntelligenceEnterprise Energy Management using a Linked Dataspace for Energy Intelligence
Enterprise Energy Management using a Linked Dataspace for Energy IntelligenceEdward Curry
 

Was ist angesagt? (20)

Smarter Computing in a New Era of IT - Dr. Gururaj Rao
Smarter Computing in a New Era of IT - Dr. Gururaj RaoSmarter Computing in a New Era of IT - Dr. Gururaj Rao
Smarter Computing in a New Era of IT - Dr. Gururaj Rao
 
Tackling big data with hadoop and open source integration
Tackling big data with hadoop and open source integrationTackling big data with hadoop and open source integration
Tackling big data with hadoop and open source integration
 
Unlocking Value in the Fragmented World of Big Data Analytics (POV Paper)
Unlocking Value in the Fragmented World of Big Data Analytics (POV Paper)Unlocking Value in the Fragmented World of Big Data Analytics (POV Paper)
Unlocking Value in the Fragmented World of Big Data Analytics (POV Paper)
 
Big data datacrunch
Big data datacrunchBig data datacrunch
Big data datacrunch
 
Dr. Shahbaz Ali, CEO at Tarmin - Business Transformation in the Age of Big Data
Dr. Shahbaz Ali, CEO at Tarmin - Business Transformation in the Age of Big DataDr. Shahbaz Ali, CEO at Tarmin - Business Transformation in the Age of Big Data
Dr. Shahbaz Ali, CEO at Tarmin - Business Transformation in the Age of Big Data
 
Quantum Leap in Open Source Collaboration
Quantum Leap in Open Source CollaborationQuantum Leap in Open Source Collaboration
Quantum Leap in Open Source Collaboration
 
Towards Expertise Modelling for Routing Data Cleaning Tasks within a Communit...
Towards Expertise Modelling for Routing Data Cleaning Tasks within a Communit...Towards Expertise Modelling for Routing Data Cleaning Tasks within a Communit...
Towards Expertise Modelling for Routing Data Cleaning Tasks within a Communit...
 
BI Forum 2009 - BI Mega Trends
BI Forum 2009 - BI Mega TrendsBI Forum 2009 - BI Mega Trends
BI Forum 2009 - BI Mega Trends
 
Prc open stack conf aug 2012 cox v1
Prc open stack conf aug 2012 cox v1Prc open stack conf aug 2012 cox v1
Prc open stack conf aug 2012 cox v1
 
Collaboration is Happening
Collaboration is HappeningCollaboration is Happening
Collaboration is Happening
 
Big Data World Forum
Big Data World ForumBig Data World Forum
Big Data World Forum
 
Data centerconsolidation10
Data centerconsolidation10Data centerconsolidation10
Data centerconsolidation10
 
Self-service Linked Government Data
Self-service Linked Government DataSelf-service Linked Government Data
Self-service Linked Government Data
 
P&O Analytics
P&O AnalyticsP&O Analytics
P&O Analytics
 
Ib Ms Vision For A Dynamic Infrastructure
Ib Ms Vision For A Dynamic InfrastructureIb Ms Vision For A Dynamic Infrastructure
Ib Ms Vision For A Dynamic Infrastructure
 
Opening keynote gianni cooreman
Opening keynote gianni cooremanOpening keynote gianni cooreman
Opening keynote gianni cooreman
 
Empowering the Business with Agile Analytics
Empowering the Business with Agile AnalyticsEmpowering the Business with Agile Analytics
Empowering the Business with Agile Analytics
 
Externalization Trend
Externalization TrendExternalization Trend
Externalization Trend
 
Enterprise 2.0 Summit 2012 Closing Keynote - Next-Generation Ecosystems And i...
Enterprise 2.0 Summit 2012 Closing Keynote - Next-Generation Ecosystems And i...Enterprise 2.0 Summit 2012 Closing Keynote - Next-Generation Ecosystems And i...
Enterprise 2.0 Summit 2012 Closing Keynote - Next-Generation Ecosystems And i...
 
Enterprise Energy Management using a Linked Dataspace for Energy Intelligence
Enterprise Energy Management using a Linked Dataspace for Energy IntelligenceEnterprise Energy Management using a Linked Dataspace for Energy Intelligence
Enterprise Energy Management using a Linked Dataspace for Energy Intelligence
 

Andere mochten auch

Single Source & Omni-Channel: The Real Power of Big Data by David Brudenell o...
Single Source & Omni-Channel: The Real Power of Big Data by David Brudenell o...Single Source & Omni-Channel: The Real Power of Big Data by David Brudenell o...
Single Source & Omni-Channel: The Real Power of Big Data by David Brudenell o...InsightInnovation
 
Insight Innovation Challenge: Single-Source & Holistic Views Of Consumers by ...
Insight Innovation Challenge: Single-Source & Holistic Views Of Consumers by ...Insight Innovation Challenge: Single-Source & Holistic Views Of Consumers by ...
Insight Innovation Challenge: Single-Source & Holistic Views Of Consumers by ...InsightInnovation
 
Architecting Cloud Solutions
Architecting Cloud SolutionsArchitecting Cloud Solutions
Architecting Cloud SolutionsAMD
 
SCF Enterprise Advisory Council
SCF Enterprise Advisory CouncilSCF Enterprise Advisory Council
SCF Enterprise Advisory CouncilSmall Cell Forum
 
AMD Fusion Developer SUmmit
AMD Fusion Developer SUmmitAMD Fusion Developer SUmmit
AMD Fusion Developer SUmmitAMD
 
Trinity press deck 10 2 2012
Trinity press deck 10 2 2012Trinity press deck 10 2 2012
Trinity press deck 10 2 2012AMD
 
02 amd fad2010 dirk_meyer_final
02 amd fad2010 dirk_meyer_final02 amd fad2010 dirk_meyer_final
02 amd fad2010 dirk_meyer_finalAMD
 
Towards 56: Small Cell Forum's HetNet & SON roadmap
Towards 56: Small Cell Forum's HetNet & SON roadmapTowards 56: Small Cell Forum's HetNet & SON roadmap
Towards 56: Small Cell Forum's HetNet & SON roadmapSmall Cell Forum
 
Strategic Business Requirements for Master Data Management Systems
Strategic Business Requirements for Master Data Management SystemsStrategic Business Requirements for Master Data Management Systems
Strategic Business Requirements for Master Data Management SystemsBoris Otto
 
Small cell backhaul: progress, challenges and next steps
Small cell backhaul: progress, challenges and next stepsSmall cell backhaul: progress, challenges and next steps
Small cell backhaul: progress, challenges and next stepsSmall Cell Forum
 
The TCO Calculator - Estimate the True Cost of Hadoop
The TCO Calculator - Estimate the True Cost of Hadoop The TCO Calculator - Estimate the True Cost of Hadoop
The TCO Calculator - Estimate the True Cost of Hadoop MapR Technologies
 
SD-WAN: Why should you care?
SD-WAN: Why should you care?SD-WAN: Why should you care?
SD-WAN: Why should you care?CloudSyntrix
 
True Single Customer View
True Single Customer View True Single Customer View
True Single Customer View Veer Endra
 
The In-Building Wireless Standard
The In-Building Wireless StandardThe In-Building Wireless Standard
The In-Building Wireless StandardSmall Cell Forum
 
bcc/BPF tools - Strategy, current tools, future challenges
bcc/BPF tools - Strategy, current tools, future challengesbcc/BPF tools - Strategy, current tools, future challenges
bcc/BPF tools - Strategy, current tools, future challengesIO Visor Project
 
Machine Learning Real Life Applications By Examples
Machine Learning Real Life Applications By ExamplesMachine Learning Real Life Applications By Examples
Machine Learning Real Life Applications By ExamplesMario Cartia
 
Race to Reality: The Next Billion-People Market Opportunity
Race to Reality: The Next Billion-People Market OpportunityRace to Reality: The Next Billion-People Market Opportunity
Race to Reality: The Next Billion-People Market OpportunityAMD
 
Hussein Mehanna, Engineering Director, ML Core - Facebook at MLconf ATL 2016
Hussein Mehanna, Engineering Director, ML Core - Facebook at MLconf ATL 2016Hussein Mehanna, Engineering Director, ML Core - Facebook at MLconf ATL 2016
Hussein Mehanna, Engineering Director, ML Core - Facebook at MLconf ATL 2016MLconf
 

Andere mochten auch (19)

Single Source & Omni-Channel: The Real Power of Big Data by David Brudenell o...
Single Source & Omni-Channel: The Real Power of Big Data by David Brudenell o...Single Source & Omni-Channel: The Real Power of Big Data by David Brudenell o...
Single Source & Omni-Channel: The Real Power of Big Data by David Brudenell o...
 
Insight Innovation Challenge: Single-Source & Holistic Views Of Consumers by ...
Insight Innovation Challenge: Single-Source & Holistic Views Of Consumers by ...Insight Innovation Challenge: Single-Source & Holistic Views Of Consumers by ...
Insight Innovation Challenge: Single-Source & Holistic Views Of Consumers by ...
 
Architecting Cloud Solutions
Architecting Cloud SolutionsArchitecting Cloud Solutions
Architecting Cloud Solutions
 
SCF Enterprise Advisory Council
SCF Enterprise Advisory CouncilSCF Enterprise Advisory Council
SCF Enterprise Advisory Council
 
AMD Fusion Developer SUmmit
AMD Fusion Developer SUmmitAMD Fusion Developer SUmmit
AMD Fusion Developer SUmmit
 
Trinity press deck 10 2 2012
Trinity press deck 10 2 2012Trinity press deck 10 2 2012
Trinity press deck 10 2 2012
 
02 amd fad2010 dirk_meyer_final
02 amd fad2010 dirk_meyer_final02 amd fad2010 dirk_meyer_final
02 amd fad2010 dirk_meyer_final
 
Towards 56: Small Cell Forum's HetNet & SON roadmap
Towards 56: Small Cell Forum's HetNet & SON roadmapTowards 56: Small Cell Forum's HetNet & SON roadmap
Towards 56: Small Cell Forum's HetNet & SON roadmap
 
Strategic Business Requirements for Master Data Management Systems
Strategic Business Requirements for Master Data Management SystemsStrategic Business Requirements for Master Data Management Systems
Strategic Business Requirements for Master Data Management Systems
 
Small cell backhaul: progress, challenges and next steps
Small cell backhaul: progress, challenges and next stepsSmall cell backhaul: progress, challenges and next steps
Small cell backhaul: progress, challenges and next steps
 
The TCO Calculator - Estimate the True Cost of Hadoop
The TCO Calculator - Estimate the True Cost of Hadoop The TCO Calculator - Estimate the True Cost of Hadoop
The TCO Calculator - Estimate the True Cost of Hadoop
 
SD-WAN: Why should you care?
SD-WAN: Why should you care?SD-WAN: Why should you care?
SD-WAN: Why should you care?
 
Docker 101 Checonf 2016
Docker 101 Checonf 2016Docker 101 Checonf 2016
Docker 101 Checonf 2016
 
True Single Customer View
True Single Customer View True Single Customer View
True Single Customer View
 
The In-Building Wireless Standard
The In-Building Wireless StandardThe In-Building Wireless Standard
The In-Building Wireless Standard
 
bcc/BPF tools - Strategy, current tools, future challenges
bcc/BPF tools - Strategy, current tools, future challengesbcc/BPF tools - Strategy, current tools, future challenges
bcc/BPF tools - Strategy, current tools, future challenges
 
Machine Learning Real Life Applications By Examples
Machine Learning Real Life Applications By ExamplesMachine Learning Real Life Applications By Examples
Machine Learning Real Life Applications By Examples
 
Race to Reality: The Next Billion-People Market Opportunity
Race to Reality: The Next Billion-People Market OpportunityRace to Reality: The Next Billion-People Market Opportunity
Race to Reality: The Next Billion-People Market Opportunity
 
Hussein Mehanna, Engineering Director, ML Core - Facebook at MLconf ATL 2016
Hussein Mehanna, Engineering Director, ML Core - Facebook at MLconf ATL 2016Hussein Mehanna, Engineering Director, ML Core - Facebook at MLconf ATL 2016
Hussein Mehanna, Engineering Director, ML Core - Facebook at MLconf ATL 2016
 

Ähnlich wie Cutting Big Data Down to Size with AMD and Dell

Splunk Overview
Splunk OverviewSplunk Overview
Splunk OverviewSplunk
 
Ibm big data hadoop summit 2012 james kobielus final 6-13-12(1)
Ibm big data    hadoop summit 2012 james kobielus final 6-13-12(1)Ibm big data    hadoop summit 2012 james kobielus final 6-13-12(1)
Ibm big data hadoop summit 2012 james kobielus final 6-13-12(1)Ajay Ohri
 
OSC2012: Big Data Using Open Source: Netapp Project - Technical
OSC2012: Big Data Using Open Source: Netapp Project - TechnicalOSC2012: Big Data Using Open Source: Netapp Project - Technical
OSC2012: Big Data Using Open Source: Netapp Project - TechnicalAccenture the Netherlands
 
Big data? No. Big Decisions are What You Want
Big data? No. Big Decisions are What You WantBig data? No. Big Decisions are What You Want
Big data? No. Big Decisions are What You WantStuart Miniman
 
Ibm big data ibm marriage of hadoop and data warehousing
Ibm big dataibm marriage of hadoop and data warehousingIbm big dataibm marriage of hadoop and data warehousing
Ibm big data ibm marriage of hadoop and data warehousing DataWorks Summit
 
Big Data: Beyond the "Bigness" and the Technology (webcast)
Big Data: Beyond the "Bigness" and the Technology (webcast)Big Data: Beyond the "Bigness" and the Technology (webcast)
Big Data: Beyond the "Bigness" and the Technology (webcast)Apigee | Google Cloud
 
Big Data: A Big Trap for Product Development
Big Data: A Big Trap for Product DevelopmentBig Data: A Big Trap for Product Development
Big Data: A Big Trap for Product DevelopmentStrategy 2 Market, Inc,
 
Hortonworks roadshow
Hortonworks roadshowHortonworks roadshow
Hortonworks roadshowAccenture
 
Nick Patience, Director Product Marketing & Strategy at Recommind - Big Data:...
Nick Patience, Director Product Marketing & Strategy at Recommind - Big Data:...Nick Patience, Director Product Marketing & Strategy at Recommind - Big Data:...
Nick Patience, Director Product Marketing & Strategy at Recommind - Big Data:...Global Business Events
 
Hadoop India Summit, Feb 2011 - Informatica
Hadoop India Summit, Feb 2011 - InformaticaHadoop India Summit, Feb 2011 - Informatica
Hadoop India Summit, Feb 2011 - InformaticaSanjeev Kumar
 
Hadoop's Opportunity to Power Next-Generation Architectures
Hadoop's Opportunity to Power Next-Generation ArchitecturesHadoop's Opportunity to Power Next-Generation Architectures
Hadoop's Opportunity to Power Next-Generation ArchitecturesDataWorks Summit
 
Big Data and Implications on Platform Architecture
Big Data and Implications on Platform ArchitectureBig Data and Implications on Platform Architecture
Big Data and Implications on Platform ArchitectureOdinot Stanislas
 
Big Data Meets Social Analytics - IBM Connect 2012 (CN-CC13)
Big Data Meets Social Analytics - IBM Connect 2012 (CN-CC13)Big Data Meets Social Analytics - IBM Connect 2012 (CN-CC13)
Big Data Meets Social Analytics - IBM Connect 2012 (CN-CC13)Mark Heid
 
Intel Cloud summit: Big Data by Nick Knupffer
Intel Cloud summit: Big Data by Nick KnupfferIntel Cloud summit: Big Data by Nick Knupffer
Intel Cloud summit: Big Data by Nick KnupfferIntelAPAC
 
Modernizing Your IT Infrastructure with Hadoop - Cloudera Summer Webinar Seri...
Modernizing Your IT Infrastructure with Hadoop - Cloudera Summer Webinar Seri...Modernizing Your IT Infrastructure with Hadoop - Cloudera Summer Webinar Seri...
Modernizing Your IT Infrastructure with Hadoop - Cloudera Summer Webinar Seri...Cloudera, Inc.
 

Ähnlich wie Cutting Big Data Down to Size with AMD and Dell (20)

Splunk Overview
Splunk OverviewSplunk Overview
Splunk Overview
 
Ibm big data hadoop summit 2012 james kobielus final 6-13-12(1)
Ibm big data    hadoop summit 2012 james kobielus final 6-13-12(1)Ibm big data    hadoop summit 2012 james kobielus final 6-13-12(1)
Ibm big data hadoop summit 2012 james kobielus final 6-13-12(1)
 
OSC2012: Big Data Using Open Source: Netapp Project - Technical
OSC2012: Big Data Using Open Source: Netapp Project - TechnicalOSC2012: Big Data Using Open Source: Netapp Project - Technical
OSC2012: Big Data Using Open Source: Netapp Project - Technical
 
Big data? No. Big Decisions are What You Want
Big data? No. Big Decisions are What You WantBig data? No. Big Decisions are What You Want
Big data? No. Big Decisions are What You Want
 
Ibm big data ibm marriage of hadoop and data warehousing
Ibm big dataibm marriage of hadoop and data warehousingIbm big dataibm marriage of hadoop and data warehousing
Ibm big data ibm marriage of hadoop and data warehousing
 
Big Data: Beyond the "Bigness" and the Technology (webcast)
Big Data: Beyond the "Bigness" and the Technology (webcast)Big Data: Beyond the "Bigness" and the Technology (webcast)
Big Data: Beyond the "Bigness" and the Technology (webcast)
 
The New Enterprise Data Platform
The New Enterprise Data PlatformThe New Enterprise Data Platform
The New Enterprise Data Platform
 
Big Data: A Big Trap for Product Development
Big Data: A Big Trap for Product DevelopmentBig Data: A Big Trap for Product Development
Big Data: A Big Trap for Product Development
 
Hortonworks roadshow
Hortonworks roadshowHortonworks roadshow
Hortonworks roadshow
 
Nick Patience, Director Product Marketing & Strategy at Recommind - Big Data:...
Nick Patience, Director Product Marketing & Strategy at Recommind - Big Data:...Nick Patience, Director Product Marketing & Strategy at Recommind - Big Data:...
Nick Patience, Director Product Marketing & Strategy at Recommind - Big Data:...
 
Introducing Splunk – The Big Data Engine
Introducing Splunk – The Big Data EngineIntroducing Splunk – The Big Data Engine
Introducing Splunk – The Big Data Engine
 
S18
S18S18
S18
 
Accelerate Return on Data
Accelerate Return on DataAccelerate Return on Data
Accelerate Return on Data
 
Hadoop India Summit, Feb 2011 - Informatica
Hadoop India Summit, Feb 2011 - InformaticaHadoop India Summit, Feb 2011 - Informatica
Hadoop India Summit, Feb 2011 - Informatica
 
Hadoop's Opportunity to Power Next-Generation Architectures
Hadoop's Opportunity to Power Next-Generation ArchitecturesHadoop's Opportunity to Power Next-Generation Architectures
Hadoop's Opportunity to Power Next-Generation Architectures
 
Big Data and Implications on Platform Architecture
Big Data and Implications on Platform ArchitectureBig Data and Implications on Platform Architecture
Big Data and Implications on Platform Architecture
 
Big Data Meets Social Analytics - IBM Connect 2012 (CN-CC13)
Big Data Meets Social Analytics - IBM Connect 2012 (CN-CC13)Big Data Meets Social Analytics - IBM Connect 2012 (CN-CC13)
Big Data Meets Social Analytics - IBM Connect 2012 (CN-CC13)
 
Kurukshetra - Big Data
Kurukshetra - Big DataKurukshetra - Big Data
Kurukshetra - Big Data
 
Intel Cloud summit: Big Data by Nick Knupffer
Intel Cloud summit: Big Data by Nick KnupfferIntel Cloud summit: Big Data by Nick Knupffer
Intel Cloud summit: Big Data by Nick Knupffer
 
Modernizing Your IT Infrastructure with Hadoop - Cloudera Summer Webinar Seri...
Modernizing Your IT Infrastructure with Hadoop - Cloudera Summer Webinar Seri...Modernizing Your IT Infrastructure with Hadoop - Cloudera Summer Webinar Seri...
Modernizing Your IT Infrastructure with Hadoop - Cloudera Summer Webinar Seri...
 

Mehr von AMD

“Zen 3”: AMD 2nd Generation 7nm x86-64 Microprocessor Core
“Zen 3”: AMD 2nd Generation 7nm x86-64 Microprocessor Core“Zen 3”: AMD 2nd Generation 7nm x86-64 Microprocessor Core
“Zen 3”: AMD 2nd Generation 7nm x86-64 Microprocessor CoreAMD
 
Heterogeneous Integration with 3D Packaging
Heterogeneous Integration with 3D PackagingHeterogeneous Integration with 3D Packaging
Heterogeneous Integration with 3D PackagingAMD
 
3D V-Cache
3D V-Cache 3D V-Cache
3D V-Cache AMD
 
AMD EPYC Family World Record Performance Summary Mar 2022
AMD EPYC Family World Record Performance Summary Mar 2022AMD EPYC Family World Record Performance Summary Mar 2022
AMD EPYC Family World Record Performance Summary Mar 2022AMD
 
AMD EPYC Family of Processors World Record
AMD EPYC Family of Processors World RecordAMD EPYC Family of Processors World Record
AMD EPYC Family of Processors World RecordAMD
 
AMD EPYC Family of Processors World Record
AMD EPYC Family of Processors World RecordAMD EPYC Family of Processors World Record
AMD EPYC Family of Processors World RecordAMD
 
AMD EPYC World Records
AMD EPYC World RecordsAMD EPYC World Records
AMD EPYC World RecordsAMD
 
AMD: Where Gaming Begins
AMD: Where Gaming BeginsAMD: Where Gaming Begins
AMD: Where Gaming BeginsAMD
 
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APU
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APUHot Chips: AMD Next Gen 7nm Ryzen 4000 APU
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APUAMD
 
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APU
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APUHot Chips: AMD Next Gen 7nm Ryzen 4000 APU
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APUAMD
 
AMD EPYC 7002 World Records
AMD EPYC 7002 World RecordsAMD EPYC 7002 World Records
AMD EPYC 7002 World RecordsAMD
 
AMD EPYC 7002 World Records
AMD EPYC 7002 World RecordsAMD EPYC 7002 World Records
AMD EPYC 7002 World RecordsAMD
 
Zen 2: The AMD 7nm Energy-Efficient High-Performance x86-64 Microprocessor Core
Zen 2: The AMD 7nm Energy-Efficient High-Performance x86-64 Microprocessor CoreZen 2: The AMD 7nm Energy-Efficient High-Performance x86-64 Microprocessor Core
Zen 2: The AMD 7nm Energy-Efficient High-Performance x86-64 Microprocessor CoreAMD
 
AMD Radeon™ RX 5700 Series 7nm Energy-Efficient High-Performance GPUs
AMD Radeon™ RX 5700 Series 7nm Energy-Efficient High-Performance GPUsAMD Radeon™ RX 5700 Series 7nm Energy-Efficient High-Performance GPUs
AMD Radeon™ RX 5700 Series 7nm Energy-Efficient High-Performance GPUsAMD
 
AMD Chiplet Architecture for High-Performance Server and Desktop Products
AMD Chiplet Architecture for High-Performance Server and Desktop ProductsAMD Chiplet Architecture for High-Performance Server and Desktop Products
AMD Chiplet Architecture for High-Performance Server and Desktop ProductsAMD
 
AMD EPYC 100 World Records and Counting
AMD EPYC 100 World Records and CountingAMD EPYC 100 World Records and Counting
AMD EPYC 100 World Records and CountingAMD
 
AMD EPYC 7002 Launch World Records
AMD EPYC 7002 Launch World RecordsAMD EPYC 7002 Launch World Records
AMD EPYC 7002 Launch World RecordsAMD
 
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance ComputingDelivering the Future of High-Performance Computing
Delivering the Future of High-Performance ComputingAMD
 
7nm "Navi" GPU - A GPU Built For Performance
7nm "Navi" GPU - A GPU Built For Performance 7nm "Navi" GPU - A GPU Built For Performance
7nm "Navi" GPU - A GPU Built For Performance AMD
 
The Path to "Zen 2"
The Path to "Zen 2"The Path to "Zen 2"
The Path to "Zen 2"AMD
 

Mehr von AMD (20)

“Zen 3”: AMD 2nd Generation 7nm x86-64 Microprocessor Core
“Zen 3”: AMD 2nd Generation 7nm x86-64 Microprocessor Core“Zen 3”: AMD 2nd Generation 7nm x86-64 Microprocessor Core
“Zen 3”: AMD 2nd Generation 7nm x86-64 Microprocessor Core
 
Heterogeneous Integration with 3D Packaging
Heterogeneous Integration with 3D PackagingHeterogeneous Integration with 3D Packaging
Heterogeneous Integration with 3D Packaging
 
3D V-Cache
3D V-Cache 3D V-Cache
3D V-Cache
 
AMD EPYC Family World Record Performance Summary Mar 2022
AMD EPYC Family World Record Performance Summary Mar 2022AMD EPYC Family World Record Performance Summary Mar 2022
AMD EPYC Family World Record Performance Summary Mar 2022
 
AMD EPYC Family of Processors World Record
AMD EPYC Family of Processors World RecordAMD EPYC Family of Processors World Record
AMD EPYC Family of Processors World Record
 
AMD EPYC Family of Processors World Record
AMD EPYC Family of Processors World RecordAMD EPYC Family of Processors World Record
AMD EPYC Family of Processors World Record
 
AMD EPYC World Records
AMD EPYC World RecordsAMD EPYC World Records
AMD EPYC World Records
 
AMD: Where Gaming Begins
AMD: Where Gaming BeginsAMD: Where Gaming Begins
AMD: Where Gaming Begins
 
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APU
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APUHot Chips: AMD Next Gen 7nm Ryzen 4000 APU
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APU
 
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APU
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APUHot Chips: AMD Next Gen 7nm Ryzen 4000 APU
Hot Chips: AMD Next Gen 7nm Ryzen 4000 APU
 
AMD EPYC 7002 World Records
AMD EPYC 7002 World RecordsAMD EPYC 7002 World Records
AMD EPYC 7002 World Records
 
AMD EPYC 7002 World Records
AMD EPYC 7002 World RecordsAMD EPYC 7002 World Records
AMD EPYC 7002 World Records
 
Zen 2: The AMD 7nm Energy-Efficient High-Performance x86-64 Microprocessor Core
Zen 2: The AMD 7nm Energy-Efficient High-Performance x86-64 Microprocessor CoreZen 2: The AMD 7nm Energy-Efficient High-Performance x86-64 Microprocessor Core
Zen 2: The AMD 7nm Energy-Efficient High-Performance x86-64 Microprocessor Core
 
AMD Radeon™ RX 5700 Series 7nm Energy-Efficient High-Performance GPUs
AMD Radeon™ RX 5700 Series 7nm Energy-Efficient High-Performance GPUsAMD Radeon™ RX 5700 Series 7nm Energy-Efficient High-Performance GPUs
AMD Radeon™ RX 5700 Series 7nm Energy-Efficient High-Performance GPUs
 
AMD Chiplet Architecture for High-Performance Server and Desktop Products
AMD Chiplet Architecture for High-Performance Server and Desktop ProductsAMD Chiplet Architecture for High-Performance Server and Desktop Products
AMD Chiplet Architecture for High-Performance Server and Desktop Products
 
AMD EPYC 100 World Records and Counting
AMD EPYC 100 World Records and CountingAMD EPYC 100 World Records and Counting
AMD EPYC 100 World Records and Counting
 
AMD EPYC 7002 Launch World Records
AMD EPYC 7002 Launch World RecordsAMD EPYC 7002 Launch World Records
AMD EPYC 7002 Launch World Records
 
Delivering the Future of High-Performance Computing
Delivering the Future of High-Performance ComputingDelivering the Future of High-Performance Computing
Delivering the Future of High-Performance Computing
 
7nm "Navi" GPU - A GPU Built For Performance
7nm "Navi" GPU - A GPU Built For Performance 7nm "Navi" GPU - A GPU Built For Performance
7nm "Navi" GPU - A GPU Built For Performance
 
The Path to "Zen 2"
The Path to "Zen 2"The Path to "Zen 2"
The Path to "Zen 2"
 

Cutting Big Data Down to Size with AMD and Dell

  • 1. COST-EFFECTIVE SOLUTIONS FOR BIG DATA WITH HADOOP Matt Kimball Product Marketing Manager Dell World 2012 Austin, Texas December 13, 2012
  • 2. WHAT‘S BIG DATA? BIG DATA Data sets sized beyond the ability of commonly used tools to capture, manage, and process the data within a tolerable time. BIG DATA Technologies A new generation of technologies designed to economically extract value from very large volumes of a wide variety of data. Definitions derived from Gartner, IDC, and NIST documents 2 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 3. DROWNING IN DATA? Proliferation of Increasing data unstructured complexity data Growing Surge in numbers of external data volumes 2.5 streams quintillion bytes of data created daily* 3 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 4. WHAT‘S HADOOP? Framework for distributed processing of large data sets across clusters of computers using a simple programming model • Scalable, fault-tolerant, distributed system for data storage and processing • Foundation of rapidly growing ecosystem of supporting projects • Created by developers to handle intractable problem (Cloudera, Ebay, Facebook, Hortonworks, Linkedin, Twitter, and Yahoo!) • Open source project hosted under the Apache Software Foundation http://hadoop.apache.org/ 4 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 5. HOW IS HADOOP BEING USED? INDUSTRY TERM VERTICAL INDUSTRY TERM Social network analysis Web Clickstream sessionization PROCESSING Media ANALYTICS ADVANCED Content optimization Clickstream sessionization DATA Network analytics Telco Mediation Loyalty and promotions analysis Retail Data factory Fraud analysis Financial Trade reconciliation Entity analysis Federal SIGINT Source: Cloudera 5 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 6. WHO IS USING, CONTRIBUTING, AND BUILDING FOR HADOOP? 10Gen Concurrent Klout SnapLogic Accela Datameer LexisNexis StackIQ Adobe eBay LinkedIn StumbleUpon Alibaba.com eHarmony Microsoft SunGard Amazon.com EMC NetApp Talend AOL EnterpriseDB Netflix Telenav AT&T Google Nokia Teradata Baidu Groupon NTT Communications The New York Times Bank of America Hortonworks Orbitz Trend Micro bitly Hulu Pandora Twitter Booz Allen Hamilton IBM Pervasive Tynt CBS IIIT Hyderabad Rackspace Hosting Vertica CBS Interactive Intel Samsung Visa Cisco Jaspersoft SGI VoltDB CNET JPMorgan Chase Sky Yahoo Source: Hortonworks And of course AMD! http://wiki.apache.org/hadoop/PoweredBy 6 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 7. WHY WAS HADOOP CREATED? Exploding data volumes Dramatic changes in enterprise data LEADS TO and types management DIGITAL CONTENT OPERATIONAL • Extract more value from NEW WEB DATA more data more cost OPPORTUNITY LOGS FOR USING effectively with greater SOCIAL FILES MEDIA SMART CREATED DATA flexibility GRIDS • Deep analysis TRANSACTIONAL HARD PROBLEMS • Exhaustive and detailed DATA • Sophisticated algorithms AD • Quick results IMPRESSIONS R&D DATA • Any type any kind BIG DATA • From any source • Structured and unstructured • It’s difficult to handle data this diverse at this scale • At scale • Traditional platforms can’t keep pace Source: Cloudera 7 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 8. TRENDS AROUND BIG DATA Enterprise and Government organizations faced with analyzing large amounts of unstructured data. Big Data Trends: • Low acquisition costs • Servers that can effectively parallelize workloads • Cost effectively scale out Hadoop systems to meet expanding data processing and analysis needs • Compute density to handle large amounts of data Data Driven Business Decisions and Actions 8 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 9. AMD AND DELL FOR YOUR HADOOP CLUSTERS CPU • Medium clock speeds Core Density for Parallelism Form Factor • Two sockets Three P Equilibrium Power • Concern as cluster grows Scale Out Capability RAM • 48GB to keep CPUs busy Manageability Storage • 1 spindle per core – SATA Cost Per Node Network • 1GB growing to InfiniBand™ 10 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 10. AMD OPTERON™ SERIES PROCESSORS FOR HADOOP LOWEST TCO FOR DELL CUSTOMERS CapEx Hadoop Deployment Scenario OpEx TCO 24 TB Data I/O Bound 10 Users Core Count C6105 AMD 4274HE Scale Out Perf 64GB RAM 12 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 11. AMD HADOOP SCENARIO TCO $800,000 $605,120 $705,040 $401,500 Up to 34% Lower $600,000 CapEx Hardware Costs11 $422,842 Dell C6105 $400,000 HP SL230s Gen8 Up to 40% Lower* $200,000 $87,920 CapEx $0 $12,000 $18,342 $3,000 $0 Server HW cost Server SW cost Rack switch cost Rack + install cost CapEx $183,977 $200,000 OpEx $142,267 $105,195 Up to 23% Lower $150,000 $100,000 $24,479 Dell C6105 OpEx* $19,583 $34,720 HP DL360e $50,000 $15,773 $12,619 $8,680 Gen8 $0 Server power Overhead Mgmt OH/year Footprint cost OpEx/yr cost/year power cost/year Up to 32% lower 3 year TCO - $407,329 savings* * See slide titled ” Substantiation | Hadoop TCO Calculations” for assumptions and calculations 13 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 12. HADOOP CONFIGURATION COMPARISON DELL POWEREDGE C6105 VS HP DL360E GEN8 Dell™ PowerEdge™ C6105 (4 nodes) HP DL360e Gen8 CPU / Node AMD Opteron™ 4274HE (2) CPU Intel Xeon E5-2440 (2) Memory / Node 64 GB (8x8 1333MHz) Memory 64 GB (8x8 1333MHz) Storage / Node 3x 2TB SATA Storage 3 x 2TB LFF SATA Network / Node 1GBe Network 1GBe PSU / Node 1100W PSU 750W 80 Servers 1 42U rack (C6105) vs. 4 42U racks (DL360e Gen8) Mellanox SX2016 10GBe switches: 2 for C6105 vs. 8 for DL360e Gen8 Hadoop Projects Software 14 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 13. EUROPEAN AEROSPACE PROJECT HADOOP SYSTEM Astronomy Mission Mapping Our Galaxy • 4 service nodes • Cluster infrastructure – 5 Nodes • 94 x C6105 – 742 Sockets • 4x MB w/ 2x 4274HE, 24GB, 3x 3TB • Dell|Force10 + Dell switches 2013 Production 15 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 14. DELL CLOUDERA SOLUTIONS The Dell | Hadoop Big Data Solution • Cloudera’s Distribution for Hadoop • Dell PowerEdge™ C hardware, including the C6105 and C6145 http://dell.cloudera.com/ http://www.mellanox.com/webinars/Accelerating-Big-Data-Over-Hadoop-With-InfiniBand/ 16 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 15. THANK YOU For more information on AMD Big Data solutions, please visit; www.AMD.com/BigData 17 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 16. DISCLAIMER The information presented in this document is for informational purposes only and may contain technical inaccuracies, omissions and typographical errors. The information contained herein is subject to change and may be rendered inaccurate for many reasons, including but not limited to product and roadmap changes, component and motherboard version changes, new model and/or product releases, product differences between differing manufacturers, software changes, BIOS flashes, firmware upgrades, or the like. AMD assumes no obligation to update or otherwise correct or revise this information. However, AMD reserves the right to revise this information and to make changes from time to time to the content hereof without obligation of AMD to notify any person of such revisions or changes. AMD MAKES NO REPRESENTATIONS OR WARRANTIES WITH RESPECT TO THE CONTENTS HEREOF AND ASSUMES NO RESPONSIBILITY FOR ANY INACCURACIES, ERRORS OR OMISSIONS THAT MAY APPEAR IN THIS INFORMATION. AMD SPECIFICALLY DISCLAIMS ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR ANY PARTICULAR PURPOSE. IN NO EVENT WILL AMD BE LIABLE TO ANY PERSON FOR ANY DIRECT, INDIRECT, SPECIAL OR OTHER CONSEQUENTIAL DAMAGES ARISING FROM THE USE OF ANY INFORMATION CONTAINED HEREIN, EVEN IF AMD IS EXPRESSLY ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. ATTRIBUTION © 2012 Advanced Micro Devices, Inc. All rights reserved. AMD, the AMD Arrow logo, Radeon, and combinations thereof are trademarks of Advanced Micro Devices, Inc. Other names and logos are used for informational purposes only and may be trademarks of their respective owners. 18 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 17. BACK-UP 19 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 18. REFERENCE CONFIGURATION Server: Dell™ PowerEdge™ C6105 (4-nodes) Software: CPU/node: AMD Opteron™ 4276HE (4) OS: CentOS 5.x (or Red Hat) Memory/node: 64 GB (8x8 1333MHz) Hadoop Storage/node: 3 x 2TB SATA Distribution: Cloudera Network/node: 1 GBe JAVA: Sun Oracle JVM 1.6u2x PSU/node: 1100W Monitoring: Nagios, Ganglia Other: Hadoop Projects Software Coming soon: AMD Opteron™ 4300 series processor support 20 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 19. AMD’S PILOT HADOOP ARCHITECTURE 21 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 20. HADOOP TCO CALCULATIONS 22 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 21. HADOOP FOR TELCO 4 x Motherboard system Hadoop on Each Motherboard: C6105 • 2 x4274HE Proc. • 4 x 16GB RAM • 4 x 2TB HDD 23 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 22. SUBSTANTIATION AND FOOTNOTES 1. SVR-217 - AMD Opteron™ 6300 Series processors are expected to have up to 37% higher java performance/watt than AMD Opteron 6200 Series processors.. Estimate based on preliminary measurements of java performance/watt in AMD labs as of August 2012 comparing AMD Opteron™ processor Models 6380 and 6278. SVR-217 2. SVR-168 - AMD Opteron™ 6300 Series processors are expected to have up to 24% higher Java performance than AMD Opteron 6200 Series processors. Estimate based on preliminary measurements of server side Java performance in AMD labs as of August 30, 2012. 1,199,838 operations per second using 2 x AMD Opteron™ processors Model 6278. 1,489,668 operations per second using 2 x AMD Opteron™ processors Model 6380. SVR-168 3. SVR-205 - AMD Opteron™ 6300 processors are expected to offer up to 8% better integer performance than AMD Opteron™ 6200 processors. Estimate based on preliminary SPECint®_rate2006 results in AMD labs as of August 2012 comparing AMD Opteron™ processor Models 6380 and 6278. 4. SVR-216 - AMD Opteron™ 6300 processors are expected to offer up to 7% better floating point performance than AMD Opteron™ 6200 processors. Estimates based on preliminary SPECfp®_rate2006 measurements in AMD labs as of August 2012 comparing AMD Opteron™ processor Models 6380 and 6278. SVR-216 5. SVR-255 - The Dell R815 lowers TCO for virtualization by up to 17% versus the HP DL560 Gen8. TCO calculation based on virtualization scenario with 1600 physical servers consolidated down to 40 with each physical server running 40 virtualized machines (12GB RAM each). This is supported by two racks of (20) servers with two Mellanox SX1016 10GBe 64-port switches. The cost per rack with installation is $3,000 based on http://www.intel.com/technology/eep/datacenter.pdf as of 9/10/12 and (2) Mellanox SX1016 10GBe 64-port switches cost $10,990 each and have a typical power consumption of 113 watts as of 9/10/12 at http://www.colfaxdirect.com/store/pc/viewPrd.asp?idproduct=1377. Power costs based on $0.10 per kWhr and using 976 watts per for the Dell R815 server and 588 watts for the HP DL560p Gen8 server according to the Dell Energy Smart Solution Advisor (ESSA) tool with a memory-intensive workload set at 100% CPU loading and an input voltage of 220 volts as well as the HP Power Advisor Tool set at 100% CPU loading and an input voltage of 220 volts as of 9/24/12 based on the configurations below. The PUE is 1.8. Assuming 28 square feet per rack according to http://www.energystar.gov/ia/partners/prod_development/downloads/Additional_FAQs_LightingCooling_Racks.pdf as of 9/11/12, and the average fully burdened cost of real estate at $310 per square foot per year based on http://www.vmware.com/technology/whyvmware/calculator/index.php as of 9/10/12. Management overhead yearly cost per server based on $81k salary with 1.25 multiplier for a fully loaded expense (benefits, taxes, etc) of $101,250. $1315 per server derived from dividing the fully loaded expense by 77 servers being managed by IT administrator, based on page 4 at http://www.vmware.com/files/pdf/vmware-solution-opex-reducing-opex-wp-en.pdf and http://www.indeed.com/salary/Windows-Server- Administrator.html. VMware vSphere license is $3495 per CPU as of 9/25/12 at www.vmware.com. R815 with four AMD Opteron™ processor Model 6274, 512GB RAM, 2 x 1TB SATA drives, two 10GBe NICs, and 3yr base warranty cost $20140 as of 9/24/12 at www.dell.com. HP DL560p Gen8 with four Intel Xeon processor Model E5-4620, 512GB RAM, 2 x 1TB SATA drives, two 10GBe NICs, and 3yr base warranty cost $30948 as of 9/24/12 at www.hp.com. The 3yr TCO using AMD Opteron™-based servers is $1,811,446. The 3yr TCO using Intel Xeon-based server is $2,170,350. Assuming customer already has agreement for client license and run time license, so not including Windows Server license. Actual results will vary. SVR-255 24 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 23. SUBSTANTIATION AND FOOTNOTES, CONTINUED 6. SVR-254 - The Dell R815 offers up to a 35% lower server hardware acquisition cost than the HP DL560 Gen8 for virtualization configurations. R815 with four AMD Opteron™ processor Model 6274, 512GB RAM, 2 x 1TB SATA drives, two 10GBe NICs, and 3yr base warranty cost $20140 as of 9/24/12 at www.dell.com. HP DL560 Gen8 with four Intel Xeon processor Model E5-4620, 512GB RAM, 2 x 1TB SATA drives, two 10GBe NICs, and 3yr base warranty cost $30948 as of 9/24/12 at www.hp.com. SVR-254 7. SVR-213 - Virtualized AMD-based servers deliver up to 12% higher performance and up to 23% better price performance than the competition. Testing with VMware ESX in today’s typical virtualized environments consisting of mid-range processor and memory configurations, reveals that AMD Opteron processors lead in both price as well as price/performance versus Intel-based systems. Virtualized AMD-based servers deliver up to 12% higher performance and up to 23% better price performance than the competition. Test run VMware ESX with DVD Store, an open source performance tool that includes web and database servers to emulate an ecommerce environment.. Servers running increasing number of VMs from 2 to 32 with a fixed, utilization rate of ~25% . Systems include: 2X AMD Opteron™ processors Model 6274 in HP DL385 Gen8 server running 4,233 Operations per Minute (OPM) for 2 VMs to 1,242 OPM for 32 VMs with 96GB (12x8 GB DDR3- 1333) memory, 1x15K 500GB SAS HD, base warranty and costing $7076 for the hardware and $6990 for VMware vSphere Enterprise Plus software; 2 x Intel Xeon E5-2665 processors in a HP DL380p Gen 8 server running 3,769 OPM for 2 VMs to 1,178 OPM for 32 VMs, 96GB (12x8 GB DDR3-1333) memory, 1x15K 500GB SAS HD, base warranty and costing $9,338 for hardware and $6990 for VMware vSphere Enterprise Plus software . System prices as of 7/5/2012 http://www.hp.com and software as of 7/5/2012 http://vmware.com. SVR-213 8. SVR-303 - The Dell C6145 lowers TCO for VDI by up to 28% versus the HP SL230s Gen8. For the Dell C6145, the TCO calculation is based on a VDI scenario with 8000 users, 200 VDI sessions per server, and 2.5GB physical RAM per session, which is supported by one rack of (20) two-node 4P servers and two Mellanox SX1016 10GBe 64-port switches. For the HP SL230s Gen8, the TCO calculation is based on a VDI scenario with 8000 users, 100 VDI sessions per server, and 2.5GB physical RAM per session, which is supported by one rack of (80) 2P servers and two Mellanox SX1016 10GBe 64- port switches. The cost per rack with installation is $3,000 based on http://www.intel.com/technology/eep/datacenter.pdf as of 9/10/12 and (2) Mellanox SX1016 10GBe 64-port switches cost $10,990 each and have a typical power consumption of 113 watts as of 9/10/12 at http://www.colfaxdirect.com/store/pc/viewPrd.asp?idproduct=1377. Power costs based on $0.10 per kWhr and using 1650 watts per for the Dell C6145 server and 417 watts for the HP SL230s server plus 182 watts for the C6500 enclosure according to the Dell Energy Smart Solution Advisor (ESSA) tool with a memory-intensive workload set at 100% CPU loading and an input voltage of 220 volts as well as the HP Power Advisor Tool set at 100% CPU loading and an input voltage of 220 volts as of 9/26/12 based on the configurations below. The PUE is 1.8. Assuming 28 square feet per rack according to http://www.energystar.gov/ia/partners/prod_development/downloads/Additional_FAQs_LightingCooling_Racks.pdf as of 9/11/12, and the average fully burdened cost of real estate at $310 per square foot per year based on http://www.vmware.com/technology/whyvmware/calculator/index.php as of 9/10/12. Management overhead yearly cost per server based on $81k salary with 1.25 multiplier for a fully loaded expense (benefits, taxes, etc) of $101,250. $1315 per server derived from dividing the fully loaded expense by 77 servers being managed by IT administrator, based on page 4 at http://www.vmware.com/files/pdf/vmware-solution-opex-reducing-opex-wp-en.pdf and http://www.indeed.com/salary/Windows-Server-Administrator.html. XenDesktop VDI software costs $134.75 per VDI session as of 9/10/12 provided by Sales at Insight. Dell C6145 with two server nodes of four AMD Opteron™ processor Model 6274, 512GB (32x16 1600MHz) RAM, 2 x 1TB SATA drives, one 10GBe NICs, and 3yr 24x7 warranty cost $43978 as of 10/4/12 according to Dell Sales. HP SL230s Gen8 with two Intel Xeon processor Model E5-2665, 256GB (16x16 1600MHz) RAM, 2 x 1TB SATA drives, one 10GBe NICs, and 3yr 24x7 warranty cost $19669 plus $2008 for the HP C6500 enclosure as of 10/4/12 at www.hp.com. The 3yr TCO using AMD Opteron™-based servers is $2,323,545. The 3yr TCO using Intel Xeon-based servers is $3,205,690. Assuming customer already has agreement for client license and run time license, so not including Windows Server license. Actual results will vary. SVR-303 25 | Cutting Big Data Down to Size | Dell World – December 2012 |
  • 24. SUBSTANTIATION AND FOOTNOTES, CONTINUED 9. SVR-302 - The Dell C6145 offers up to a 44% lower server hardware acquisition cost than the HP SL230s Gen8 for VDI configurations. Dell C6145 with two server nodes of four AMD Opteron™ processor Model 6274, 512GB (32x16 1600MHz) RAM, 2 x 1TB SATA drives, one 10GBe NICs, and 3yr 24x7 warranty cost $43978 as of 10/4/12 according to Dell Sales. HP SL230s Gen8 with two Intel Xeon processor Model E5-2665, 256GB (16x16 1600MHz) RAM, 2 x 1TB SATA drives, one 10GBe NICs, and 3yr 24x7 warranty cost $19669 as of 10/4/12 at www.hp.com. SVR-302 10. SVR-301 - The Dell C6105 lowers TCO for Hadoop by 32% versus the HP DL360e Gen8. TCO calculation based on Hadoop scenario with 24TB dataset being analyzed by 10 users and being processed by an I/O bound application. This is supported by one rack of 79 Dell C6105 slave nodes and 1 Dell C6105 master node along with two Dell Force10 S60 High-Performance 1/10 GbE switches as well as four racks of 79 HP DL360e Gen8 slave nodes and 1 HP DL360e Gen8 master node along with two Mellanox SX1016 10GBe 64-port switches per rack. The cost per rack with installation is $3,000 based on http://www.intel.com/technology/eep/datacenter.pdf as of 9/10/12. Mellanox SX1016 10GBe 64-port switches cost $10,990 each and have a typical power consumption of 113 watts as of 9/10/12 at http://www.colfaxdirect.com/store/pc/viewPrd.asp?idproduct=1377. Dell Force10 S60 High-Performance 1/10 GbE switches cost $9171 each and have a typical power consumption of 156 watts as of 9/26/12 at www.dell.com and http://www.networkcomputing.com/next-gen-network-tech-center/force10s-deep-buffer-switch/229501471. Power costs based on $0.10 per kWhr and using 889 watts per for the Dell C6105 with four server nodes in one system and 338 watts for the HP DL360e Gen8 server with a transactional workload set at 100% CPU loading according to the Dell Energy Smart Solution Advisor (ESSA) tool with a transactional workload set at 100% CPU loading and an input voltage of 220 volts and the HP Power Advisor tool set at 100% CPU loading and an input voltage of 220 volts as of 9/26/12 based on the configurations below. The PUE is 1.8. Assuming 28 square feet per rack according to http://www.energystar.gov/ia/partners/prod_development/downloads/Additional_FAQs_LightingCooling_Racks.pdf as of 9/11/12, and the average fully burdened cost of real estate at $310 per square foot per year based on http://www.vmware.com/technology/whyvmware/calculator/index.php as of 9/10/12. Management overhead yearly cost per server based on $81k salary with 1.25 multiplier for a fully loaded expense (benefits, taxes, etc) of $101,250. $1315 per server node derived from dividing the fully loaded expense by 77 servers being managed by IT administrator, based on page 4 at http://www.vmware.com/files/pdf/vmware-solution-opex-reducing-opex-wp-en.pdf and http://www.indeed.com/salary/Windows-Server-Administrator.html. Open Source software is being used so there is no cost. Each of the four Dell C6105 server nodes has two AMD Opteron™ processor Model 4274HE, 64GB RAM, 3 x 2TB SATA drives, 1GBe, and 3yr base warranty with a total system cost of $20075 ($5019 each node) as of 10/3/12 from Dell Sales. HP DL360e with two Intel Xeon processor Model E5-2440, 64GB RAM, 3 x 2TB SATA drives, 1GBe, and 3yr base warranty cost $7564 as of 9/21/12 at www.hp.com. The 3yr TCO using AMD Opteron™-based servers is $849,643. The 3yr TCO using Intel Xeon-based servers is $1,256,972. Actual results will vary. SVR-301. 11. SVR-300 - The Dell C6105 offers up to a 34% lower scale-out server hardware acquistion cost than the HP DL360e Gen8 for Hadoop configurations. Each of the four Dell C6105 server nodes has two AMD Opteron™ processor Model 4274HE, 64GB RAM, 3 x 2TB SATA drives, 1GBe, and 3yr base warranty with a total system cost of $20075 ($5019 each node) as of 10/3/12 from Dell Sales. HP DL360e with two Intel Xeon processor Model E5-2440, 64GB RAM, 3 x 2TB SATA drives, 1GBe, and 3yr base warranty cost $7564 as of 9/21/12 at www.hp.com. SVR-300 26 | Cutting Big Data Down to Size | Dell World – December 2012 |