Your SlideShare is downloading. ×
0
Answers to the Scaling Challenge: A Case Study with Answers.com on Scaling with Memcached and MySQL<br />Presenters:<br />...
Agenda<br />Intro<br />Answers.com background<br />Company, technology<br />Scaling challenges<br />Gear6 Perspectives<br ...
Answers to the Scaling Challenge:<br />A Case Study With Answers.com on Scaling With Memcached and MySQL<br />April 14, 20...
Overview<br />The world’s leading Q&A site<br />4<br />
Overview<br />Answers.com<br /><ul><li>Combines power of WikiAnswers® community-driven content with expert licensed conten...
Users can ask anything and automatically receive the best available answer.
WikiAnswers</li></ul>Community generated social knowledge Q&A platform, leveraging wiki-based technologies - answers impro...
Answers.com Sites<br />Rank in top Web properties<br /><ul><li>#18in the U.S. (02/2010)(1)
#31worldwide (02/2010)(1)</li></ul>Unique monthly visitors<br /><ul><li>50 million in the U.S. (02/2010)(1)
72 million worldwide (02/2010)(1)</li></ul>WikiAnswers domain fastest growing in U.S.:<br /><ul><li>2007 unique visitors g...
2008 unique visitors grew 154%(#1 of top 200 Web domains) (1)
2009 unique visitors grew 74% (#2 of top 50 Web domains) (1)</li></ul>6<br />(1) Source: comScore – Hybrid Measurement Met...
Answers.com Sites<br />7<br />50MM Unique Monthly U.S. Visitors = #18<br />Feb 2010<br />Source: comScore – Hybrid Measure...
ReferenceAnswers<br />8<br />
WikiAnswers: Q&A the Wiki Way<br />9<br />
Some key infrastructure elements<br />2 Data Centers – Active/Active<br />HP Blade Servers (BL460) and 2U Servers (DL380)<...
Software<br />11<br />Linux – mainly CentOS 5.x/64bit<br />Apache/JBOSS<br />Lucene / Solr<br />Memcached<br />MySQL 5.0.x...
Secret to Web 2.0<br />Everything runs from memory in Web 2.0.<br />Evan Weaver – Twitter<br />12<br />
Questions: How do I Scale????<br />Four answers to the scaling challenge from Answers.com:<br />Use Enterprise-grade PCI-S...
Scaling Answer #1<br />Use Enterprise-grade SSDs in your MySQL servers<br />14<br />
Using SSDs in your MySQL servers<br />If you have a large DB read clusters<br />Can improve DB server queries/sec by 900%<...
Quick plug  :-)<br />Panel:<br />How Solid-state Technologies are <br />Transforming MySQL Server Performance<br /> and th...
Scaling Answer #2<br />Enhance Memcached tier<br />17<br />
Memcached tier - before<br />10 x 16/32GB RAM Memcached Servers per Data Center<br />Divided into several clusters<br />St...
Memcached tier – Challenges<br />No redundancy	<br />Lose 25% of cache (or worse) on any server failure<br />Loss of cache...
Cache data critical to performance<br />10’s of millions of pages<br />Pages are dynamic – always publishing<br />Unify da...
Multiple entries/page<br />[Illustrate mult. data entries/page]<br />21<br />
22<br />
23<br />
Cache data critical to performance<br />Multiple data sources (entries) per page<br />Each entry can have multiple links t...
Internal links for each entry</li></ul>24<br />
Lose memcached node?<br />Lose 25% cached data (or more)<br />2-20 addit. MySQL queries to retrieve metadata and data need...
Original plan to solve memcached weakness<br />Engineering team to develop memcached wrapper<br />Mirrored memcached serve...
Handle node failure
Handle cache data restore when node becomes available again</li></ul>Also: Address memcached slab issues<br />26<br />
Alternative solution – Gear6<br />Built-in redundancy function (pair of boxes)<br />Mirroring (writes *everything* to both...
Gear6 GUI config example<br />28<br />
Testing<br />Same blisteringly fast memcached performance <br /><ul><li>use of slower SSD was undetectable</li></ul>Tried ...
Value-add with Gear6 memcached<br />Improve app reliability<br />Ensure no SPOF at memcached layer<br />Scale up our datab...
Gear6 Value-add                          …cont’d<br />Insight into memcached performance<br />GUI helped us troubleshoot a...
Scaling Answer #3<br />Controlling DB Slave server clusters<br />32<br />
Controlling DB Slave server clusters<br />Maint. required sending command to Load Balancer to disable slave node in cluste...
Using Lighty to Control DB Slave clusters<br />Configlighty to make lightweight “healthcheck” call to local DB (SHOW STATU...
Lighty DB health-check<br />35<br />
Lighty script logic<br />if maint_flag = TRUE<br />		return “red”<br />elseifdbcall not good<br />		return “red”<br />else...
Scaling Answer #4<br />HW Load Balancers<br />37<br />
Hardware Load Balancers<br />Avoid making firewall your bottleneck<br />Immense flexibility [CLI, stats, ecv]<br />Quick r...
Which Hardware Load Balancers?<br />Citrix & F5<br /><ul><li>Leaders of the pack (Garter Magic Quadrant)
Amazing performance
Supremely flexible
Advanced GSLB
Expensive</li></ul>A10 Networks<br /><ul><li>Up and coming
Great performance
Decent feature set
Cost-effective (certainly for internal load-balancing)</li></ul>39<br />
40<br />Thank you.<br />Slides: http://tinyurl.com/mysqlconf2010-scaling-tips<br />
Upcoming SlideShare
Loading in...5
×

Case Study with Answers.com on Scaling with Memcached and MySQL

5,675

Published on

Presentation given at MySQL Conference 2010

Published in: Technology
1 Comment
8 Likes
Statistics
Notes
  • ssds可以极大增加性能
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
No Downloads
Views
Total Views
5,675
On Slideshare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
124
Comments
1
Likes
8
Embeds 0
No embeds

No notes for slide
  • These three are interrelated Content is Dynamic and PersonalTake a step back…Many more people and not just in the USATraffic (cisco petabyte data): measuring in petabytes per monthApplications: Dynamic and personalized
  • A lot of this population is coming into US-based social networking, gaming, media sites… distance means latency issue.Proliferation of broadband drove users to the net.USA #33 in broadband (&gt;2Mbps); #16 in average speed. Miniwatts data shows the shift to Europe and Asia in particular.World internet usage is only NOW @ 1.6b now (24% according to 50x15 site)NOW MOBILE is adding rapidly to this population explosion.Akamai state of the internet report for q1 2009 gives very insightful information of current data trends.
  • Consumer is driving the traffic growth (3x of business)&gt;40% CAGR growth (4x in 4 years)
  • How do sites make the same software stack go faster
  • Why we are here again:Scalable stack framework for LAMP and Java had/have to emerge in order to accelerate stack and provide distribution services
  • Why we are here again:Scalable stack framework for LAMP and Java had/have to emerge in order to accelerate stack and provide distribution services
  • 100ms latency introduced by traversing internet to origin site Typically +0.1 secs means 2-4% DROP in revenue
  • 100ms latency introduced by traversing internet to origin site Typically +0.1 secs means 2-4% DROP in revenue
  • Maslow&apos;s Hierarchy Of Needs
  • Increases revenue via being able to economically scale and support more members, services and transactionsScalable stack framework delivered as a network service addressable by all stack componentsOnramp information into Memory ASAPMemcached has emerged as de facto protocol to load information into memory BUT a lot more is neededMemcached ++ Focused on caching functions Advance network-based distributed services Non relational functions
  • Web 2.0 driving user &amp; traffic growth Key Apps: Entertainment, Communication, Social Networking Key Drivers: Mobile, Broadband, International, ConsumerMobile applications – 3G+ is 22% by 2010 (Mobile goes hi speed)Consumer Internet Traffic 2006–2012 This category encompasses any IP traffic that crosses the Internet and is not confined to a single service provider’s network. Peer-to-peer (P2P) traffic, still the largest share of Internet traffic today, will decrease as a percentage of overall Internet traffic. Internet video streaming and downloads are beginning to take a larger share of bandwidth, and will grow to nearly 50 percent of all consumer Internet traffic in 2012.
  • We talked about “Universal Distro” and EC2 beforeGear6 Web Cache Software requirements:Certified on HP, Dell, Rackable…Requires dedicated serverRedundant GE ports or 10 GE ports64-bit (x86 64) server with 32GB or more of RAMIDE/SATA HD with 80GB+Option: 2.5” SATA drive bays for Flash SSDs - Intel X25-E or SamsungGear6 Web Cache delivers:High availability and consistent scalable high performance50k-200k Memcached ops per sec32-300+GB Memcached space per rack unitFault toleranceComplete cache control and visibility of resourcesCost effective mechanism of scaling database and application tiers
  • We talked about “Universal Distro” and EC2 beforeGear6 Web Cache Software requirements:Certified on HP, Dell, Rackable…Requires dedicated serverRedundant GE ports or 10 GE ports64-bit (x86 64) server with 32GB or more of RAMIDE/SATA HD with 80GB+Option: 2.5” SATA drive bays for Flash SSDs - Intel X25-E or SamsungGear6 Web Cache delivers:High availability and consistent scalable high performance50k-200k Memcached ops per sec32-300+GB Memcached space per rack unitFault toleranceComplete cache control and visibility of resourcesCost effective mechanism of scaling database and application tiers
  • Transcript of "Case Study with Answers.com on Scaling with Memcached and MySQL"

    1. 1. Answers to the Scaling Challenge: A Case Study with Answers.com on Scaling with Memcached and MySQL<br />Presenters:<br /> Dan Marriott, Director - Production Operations, Answers.com<br /> Joaquín Ruiz, EVP Products, Gear6<br />
    2. 2. Agenda<br />Intro<br />Answers.com background<br />Company, technology<br />Scaling challenges<br />Gear6 Perspectives<br />Scaling needs levels: patterns<br />Technology challenges<br />Emerging technologies<br />Questions<br />
    3. 3. Answers to the Scaling Challenge:<br />A Case Study With Answers.com on Scaling With Memcached and MySQL<br />April 14, 2010<br />Dan MarriottDirector - Production Operationsdanm@answers.com<br />
    4. 4. Overview<br />The world’s leading Q&A site<br />4<br />
    5. 5. Overview<br />Answers.com<br /><ul><li>Combines power of WikiAnswers® community-driven content with expert licensed content on ReferenceAnswers™
    6. 6. Users can ask anything and automatically receive the best available answer.
    7. 7. WikiAnswers</li></ul>Community generated social knowledge Q&A platform, leveraging wiki-based technologies - answers improve over time.<br /><ul><li>ReferenceAnswers </li></ul>Content on millions of topics from over 250 licensed dictionaries and encyclopedias from leading publishers.<br />5<br />
    8. 8. Answers.com Sites<br />Rank in top Web properties<br /><ul><li>#18in the U.S. (02/2010)(1)
    9. 9. #31worldwide (02/2010)(1)</li></ul>Unique monthly visitors<br /><ul><li>50 million in the U.S. (02/2010)(1)
    10. 10. 72 million worldwide (02/2010)(1)</li></ul>WikiAnswers domain fastest growing in U.S.:<br /><ul><li>2007 unique visitors grew 573%(#1 of top 1,500 Web domains) (1)
    11. 11. 2008 unique visitors grew 154%(#1 of top 200 Web domains) (1)
    12. 12. 2009 unique visitors grew 74% (#2 of top 50 Web domains) (1)</li></ul>6<br />(1) Source: comScore – Hybrid Measurement Methodology (U.S. only) beginning August 2009<br />
    13. 13. Answers.com Sites<br />7<br />50MM Unique Monthly U.S. Visitors = #18<br />Feb 2010<br />Source: comScore – Hybrid Measurement Methodology for Answers sites beginning August 2009<br />
    14. 14. ReferenceAnswers<br />8<br />
    15. 15. WikiAnswers: Q&A the Wiki Way<br />9<br />
    16. 16. Some key infrastructure elements<br />2 Data Centers – Active/Active<br />HP Blade Servers (BL460) and 2U Servers (DL380)<br />Fusion-io SSDs in all large MySQL DB servers<br />VMware<br />HP/LeftHand SAN<br />Hardware Load-Balancers<br />10<br />
    17. 17. Software<br />11<br />Linux – mainly CentOS 5.x/64bit<br />Apache/JBOSS<br />Lucene / Solr<br />Memcached<br />MySQL 5.0.x<br />Apache-Tomcat<br />Memcached<br />MySQL 5.0.x<br />
    18. 18. Secret to Web 2.0<br />Everything runs from memory in Web 2.0.<br />Evan Weaver – Twitter<br />12<br />
    19. 19. Questions: How do I Scale????<br />Four answers to the scaling challenge from Answers.com:<br />Use Enterprise-grade PCI-SSDs in your MySQL servers<br />Enhance Memcached tier<br />Controlling DB Slave server clusters<br />Hardware Load-Balancers<br />13<br />
    20. 20. Scaling Answer #1<br />Use Enterprise-grade SSDs in your MySQL servers<br />14<br />
    21. 21. Using SSDs in your MySQL servers<br />If you have a large DB read clusters<br />Can improve DB server queries/sec by 900%<br /><ul><li>and thereby reduce # servers needed per cluster</li></ul>Can improve overall app speed<br />Use PCI-based SSDs<br /><ul><li>*Much* faster than SAS/SATA-based SSDs</li></ul>Give yourself at least 50% headroom <br /><ul><li>(to handle uncompressing copies locally on SSD)</li></ul>15<br />
    22. 22. Quick plug :-)<br />Panel:<br />How Solid-state Technologies are <br />Transforming MySQL Server Performance<br /> and the Datacenter Architectures<br />Sumeet Bansal (Fusion-io)<br />Ryan White (Cloudmark)<br />Dan Marriott (Answers.com)<br />Vadim Tkachenko (Percona Inc)<br />Jeremy Zawodny (craigslist.org) <br />Weds 5:15pm - Ballroom D<br />16<br />
    23. 23. Scaling Answer #2<br />Enhance Memcached tier<br />17<br />
    24. 24. Memcached tier - before<br />10 x 16/32GB RAM Memcached Servers per Data Center<br />Divided into several clusters<br />Striped instances across each cluster<br />10’s millions items/instance<br />18<br />
    25. 25. Memcached tier – Challenges<br />No redundancy <br />Lose 25% of cache (or worse) on any server failure<br />Loss of cache = poor performance/user experience<br />Costly (OpEx: Rackspace, power, cooling, maint, admin)<br />19<br />
    26. 26. Cache data critical to performance<br />10’s of millions of pages<br />Pages are dynamic – always publishing<br />Unify data on same topic from different data sources onto one page<br />20<br />
    27. 27. Multiple entries/page<br />[Illustrate mult. data entries/page]<br />21<br />
    28. 28. 22<br />
    29. 29. 23<br />
    30. 30. Cache data critical to performance<br />Multiple data sources (entries) per page<br />Each entry can have multiple links to other entries<br />Cache metadata:<br /><ul><li>List of entries for each page
    31. 31. Internal links for each entry</li></ul>24<br />
    32. 32. Lose memcached node?<br />Lose 25% cached data (or more)<br />2-20 addit. MySQL queries to retrieve metadata and data needed to construct page<br />Heavy load on MySQL Servers<br />Longer page load times<br />Site slows down<br />25<br />
    33. 33. Original plan to solve memcached weakness<br />Engineering team to develop memcached wrapper<br />Mirrored memcached server clusters (2x hardware)<br />6 man-months to develop:<br /><ul><li>Write to both clusters
    34. 34. Handle node failure
    35. 35. Handle cache data restore when node becomes available again</li></ul>Also: Address memcached slab issues<br />26<br />
    36. 36. Alternative solution – Gear6<br />Built-in redundancy function (pair of boxes)<br />Mirroring (writes *everything* to both nodes)<br />200GB Memcached space per box (RAM+SSD)<br />Uses standard VIP mechanism<br />Graceful failover (no impact)<br />Automatic cache-sync on node recovery<br />Auto fail-back and rebalance VIP<br />No code changes necessary<br />27<br />
    37. 37. Gear6 GUI config example<br />28<br />
    38. 38. Testing<br />Same blisteringly fast memcached performance <br /><ul><li>use of slower SSD was undetectable</li></ul>Tried to break it<br /><ul><li>yank power, kill network ports, pull SSD drive</li></ul>Try software upgrade<br /><ul><li>seamless failover/cache-sync on node-up</li></ul>Discovered vastly improved slab management<br /><ul><li>plus ability to fine-tune instance memory usage</li></ul>Scalable<br /><ul><li>can handle 60,000+ ops/sec</li></ul>29<br />
    39. 39. Value-add with Gear6 memcached<br />Improve app reliability<br />Ensure no SPOF at memcached layer<br />Scale up our database infrastructure safely<br />Significantly reduce TCO by decommissioning 20 servers<br />Save 6 man-months of custom memcached wrapper development<br />30<br />
    40. 40. Gear6 Value-add …cont’d<br />Insight into memcached performance<br />GUI helped us troubleshoot app issues<br />First class support (even at 4:00am!)<br />New in v3.0: Ability to dynamically resize memcached instances with no cache loss!!<br />31<br />
    41. 41. Scaling Answer #3<br />Controlling DB Slave server clusters<br />32<br />
    42. 42. Controlling DB Slave server clusters<br />Maint. required sending command to Load Balancer to disable slave node in cluster<br />On DB node failure, lost ‘000s queries before monitoring noticed and issued LB command<br />----------<br />Solution: Install lighty (lighttpd) on every DB server<br />33<br />
    43. 43. Using Lighty to Control DB Slave clusters<br />Configlighty to make lightweight “healthcheck” call to local DB (SHOW STATUS LIKE 'Threads_running')<br />Can set maint. flag on any DB server<br />Have LB monitor make http call to lighty (http://servername:3300/dbcheck?p=3309)<br /><ul><li>Works for multiple MySQL instances on same box</li></ul>34<br />
    44. 44. Lighty DB health-check<br />35<br />
    45. 45. Lighty script logic<br />if maint_flag = TRUE<br /> return “red”<br />elseifdbcall not good<br /> return “red”<br />else<br /> return “green”<br />end<br />36<br />
    46. 46. Scaling Answer #4<br />HW Load Balancers<br />37<br />
    47. 47. Hardware Load Balancers<br />Avoid making firewall your bottleneck<br />Immense flexibility [CLI, stats, ecv]<br />Quick reconfigs, adds/changes/deletes<br />Highly secure<br />Caching, Compression, GSLB, throttle for safety<br />38<br />
    48. 48. Which Hardware Load Balancers?<br />Citrix & F5<br /><ul><li>Leaders of the pack (Garter Magic Quadrant)
    49. 49. Amazing performance
    50. 50. Supremely flexible
    51. 51. Advanced GSLB
    52. 52. Expensive</li></ul>A10 Networks<br /><ul><li>Up and coming
    53. 53. Great performance
    54. 54. Decent feature set
    55. 55. Cost-effective (certainly for internal load-balancing)</li></ul>39<br />
    56. 56. 40<br />Thank you.<br />Slides: http://tinyurl.com/mysqlconf2010-scaling-tips<br />
    57. 57. Gear6: A few perspectives on the scaling challenge<br />
    58. 58. Web Scaling: What’s Changed<br />MASSIVE increase in:<br />Traffic<br />Population<br />Content & Applications<br />
    59. 59. Web Growth: Population<br />Forrester: 2.2 billion people online globally by 2013<br />
    60. 60. Web Growth: Traffic<br />Cisco: “Annual global IP trafficwill exceed two-thirds of a zettabyte (667 exabytes) in four years”[by 2013]<br />Source: Cisco Visual Networking Index,9 June 2009<br />
    61. 61. Web Growth: Application & Content<br />Entertainment<br />Social networking<br />Media<br />Communication<br />Community-generated content<br />Dynamic<br />Static<br />
    62. 62. Web Architectures<br />Essentially the SAME Major Building Blocks<br />Web<br />App<br />RDBMS<br />
    63. 63. Web 2.0 Architecture<br /><ul><li> Shift to Web 2.0 services puts strain on origin sites</li></ul>Web Stack<br />Net<br />Interface<br />Clients<br />Storage<br />App<br />Web<br />RDBMS<br />Internet<br />Storage Interface:<br />CDN<br />Proxy<br />LoadBalancer<br />
    64. 64. Web 2.0 Constraints<br /><ul><li> Can’t Accelerate Using CDN</li></ul>Web Stack<br />Net<br />Interface<br />Clients<br />Storage<br />App<br />Web<br />RDBMS<br />Internet<br />Storage Interface:<br />X<br />CDN<br />Proxy<br />LoadBalancer<br />
    65. 65. Web 2.0 Issue #1<br /><ul><li> LATENCY: The Web 2.0 Killer</li></ul>Web Stack<br />Net<br />Interface<br />Clients<br />Storage<br />App<br />Web<br />RDBMS<br />Internet<br />Storage Interface:<br />X<br />CDN<br />Proxy<br />LoadBalancer<br />Latency<br />0.01-0.1 secs<br />0.1-0.2 secs<br />1 secs<br />FAIL!<br />Latency: Time is the Web 2.0 killer<br />
    66. 66. Web 2.0 Needs: Data Acceleration<br />Hierarchy of Needs: Survive<br /> 1. Accelerated Data: i.e. CACHE = Survival!<br />Web Stack<br />Net<br />Interface<br />Clients<br />Storage<br />App<br />Web<br />RDBMS<br />Internet<br />Storage Interface:<br />X<br />CDN<br />Proxy<br />LoadBalancer<br />Data Cache<br />
    67. 67. Web 2.0 Needs: Data Acceleration<br />In LAMP/JAVA/Ruby World<br />Data Cache = Memcached<br />Web Stack<br />Net<br />Interface<br />Clients<br />Storage<br />App<br />Web<br />RDBMS<br />Internet<br />Storage Interface:<br />X<br />CDN<br />Proxy<br />LoadBalancer<br />Memcached<br />
    68. 68. Scaling: More Levels of Needs<br />Memcached = Survival<br />Is there more to Life?<br />Acceleration Services<br />Memory On-boarding<br />Data Management<br />Persistence?<br />Query?<br />Operations?<br />Migration/Sharing?<br />
    69. 69. JR Scratch<br />Dynamic Data Platform  Dynamic Data Services<br />Services<br />On-boarding<br />Data Management<br />Persistence<br />Query<br />Operations<br />
    70. 70. What to do<br />Accelerate<br />
    71. 71. The Gear6 Answer<br />Web Stack<br />Net<br />Interface<br />With Gear6 Web Cache:<br /><ul><li> Massively reduces latency
    72. 72. Increases revenue
    73. 73. Decreases cost, complexity and downtime</li></ul>Clients<br />Storage<br />App Servers<br />PHP, Java, Rails, C, Perl, Python…<br />Web Servers<br />Database<br />MySQL, Oracle…<br />Internet<br />Storage Interface:<br />CDN<br />Proxy<br />LoadBalancer<br />Gear6 Web Cache<br />0.01-0.1 secs<br />
    74. 74. Direction: Scale Out Platform for Dynamic Data Services<br />Web Stack<br />Net<br />Interface<br />Clients<br />Storage<br />Web Servers<br />App Servers<br />PHP, Java, Rails, C, Perl, Python<br />RDBMS<br />MySQL<br />Internet<br />Storage Interface:<br />file, block, FC, SCSI<br />CDN<br />Proxy<br />LoadBalancer<br />client cache<br />Storage cache<br />Cache Servers<br />memcached<br />Key-Value Store<br />NoSQL<br />Gear6 Dynamic<br /> Data Services<br />
    75. 75. Current Scale Out Architectures<br />Memcached Pool<br />Applications that touch disk = dead<br />App Servers<br />SQL Store<br />Large Number of Viewers/Users<br />
    76. 76. Scale Out Architectures - Next<br />KV Store<br />Applications that touch disk = dead<br />App Servers<br />Large Number of Viewers/Users<br />
    77. 77. Gear6 Web Cache<br />Gear6 Web Cache<br />Packaging:<br />Software image for pre-certified servers<br />Pre-configured appliance<br />VMware image<br />RPM & DEB packages<br />Features:<br />High performance<br />Highest-density: RAM + Flash SSD; from 32 to 384 GB / rack unit<br />Clustering, Replication & Persistence <br />Advanced Memcached management<br />Multi-tenancy<br />Benefits:<br />Scalable performance<br />High availability<br />Increased cache = Increase revenue<br />Low TCO & lower complexity than DIY<br />Consolidation<br />Memcached Support<br />Partitioned Services<br />Web Cache Platforms<br />
    78. 78. Gear6 Cloud Cache<br />Gear6 Cloud Cache<br />Cloud distribution<br />For dedicated images<br />Packaging:<br />EC2: 32 and 64-bit<br />GoGrid: 32 and 64-bit<br />Features:<br />Elastic Sizing<br />Native query support<br />Highest efficiency Memcached servers<br />Use of image’s allocated storage for Memcached<br />Auto-config and advanced reporting via GUI or REST interfaces<br />Free and paid images<br />Benefits:<br />Ease-of-use<br />High performance<br />More cached objects per MB<br />Expanded Memcached space per image<br />Memcached Support<br />Cloud Cache Platforms<br />
    1. A particular slide catching your eye?

      Clipping is a handy way to collect important slides you want to go back to later.

    ×