RavenDB 3.0 Keynote
Oren Eini
ayende@ayende.com ayende.com/blog
Hibernating Rhinos
About this conference
 Twit with
 Logistics
Free stuff
 RavenDB Coupon: ravenconf2014
 RavenDB In Action 45% discount: ravdb14cf
 Metrics beyond anything you’ve seen
 Recent Microsoft Case Study at: http://goo.gl/GTgkv6
 VIP Subscription for attendees:
 codealike.com/VIP
 Token: RavenConf2014
 Get 3 months of Codealike Premium
History
 Mid 2008 - Rhino Divan
DB Started
 Sep 2009 - RavenDB is
created
 May 2010 - RavenDB 1.0
 Nov 2010 - RavenDB 1st
production deployment
 Jan 2013 - RavenDB 2.0
 Jul 2013 - RavenDB 2.5
 Aug 2013 – 1st RavenDB
book it out
 Apr 2014 -1st RavenDB
Conference
 Jul 2014 (est) –
RavenDB 3.0 launch
RavenDB (ohloh)
 40,100 commits
 210 contributors
 2,774,921 lines of code
 estimated 797 years of effort (COCOMO model)
 first commit in September, 2009
 most recent commit about 13 hours ago
Jan 1st 2015
 RavenDB becomes self aware
RavenDB 3.0
 ~15 team members
 Some parts were started in 2011
 18 months of work (another 2/3 remaining)
 More than 600 issues
 Awesome
What did we do?!
 Voron
 OWIN / Web API
 Indexing
 Operations
 RavenFS
 JVM client API
 Spit & polish
 New studio
behold
http://www.wizards.com/dnd/images/leof_gallery/86716.jpg
It’s not about the UI
 Yes, important
 Yes, we have ~8 people on it now
 We’re a database
 http://issues.hibernatingrhinos.com
 Features, not cosmetics
Seriously, now…
 What should you be excited about?
 It isn't the feature, it is the direction…
Removing friction
Indexing
 Index deletes are async
 Index ids
 Small collection
optimization
 Fan out prevention
Operations
 No performance counters
 Additional debug endpoints
 Periodic backups bundle
full/incremental & deletes
 Explicit failover servers
 Reduced # of assemblies
 Server to server smuggling
Increasing access
Raven File system
 Tailored persistence solution
 In production since 2012
 Replicated file system
 Optimized change tracking
 Very large files
 Replacing attachments
Spit & polish
 Preserving missing properties
 Lazy async
 Single pipeline (embedded / http)
 Multiple database support for embedded
 Everything on top of OWIN / Web API
Where are we now?
 Stabilization
 http://issues.hibernatingrhinos.com
 Force new feature mode: On
 No new features going in until release
Hi, what about Voron?!
 Internal only
 Not important
!!Not so fast!
 Voron is very important
 Impl. details after lunch…
 Implications of Voron are:
 We own the entire stack
 Tailored solutions
Where are we going?
Actual
 Voron - distribution
 Log shipping
 Raft
 Polyglot persistence
solution
 RavenFS is just the
beginning
 Event aggregations, the
fallen feature
Research
 Project Corax
 Project Tempus
 Project Duco
Hackaton
 After hours
 Let us make something cool!
 Full feature, from the disk to the UI
Questions?

RavenDB 3.0 Keynote

  • 1.
    RavenDB 3.0 Keynote OrenEini ayende@ayende.com ayende.com/blog Hibernating Rhinos
  • 2.
    About this conference Twit with  Logistics
  • 3.
    Free stuff  RavenDBCoupon: ravenconf2014  RavenDB In Action 45% discount: ravdb14cf
  • 4.
     Metrics beyondanything you’ve seen  Recent Microsoft Case Study at: http://goo.gl/GTgkv6  VIP Subscription for attendees:  codealike.com/VIP  Token: RavenConf2014  Get 3 months of Codealike Premium
  • 5.
    History  Mid 2008- Rhino Divan DB Started  Sep 2009 - RavenDB is created  May 2010 - RavenDB 1.0  Nov 2010 - RavenDB 1st production deployment  Jan 2013 - RavenDB 2.0  Jul 2013 - RavenDB 2.5  Aug 2013 – 1st RavenDB book it out  Apr 2014 -1st RavenDB Conference  Jul 2014 (est) – RavenDB 3.0 launch
  • 6.
    RavenDB (ohloh)  40,100commits  210 contributors  2,774,921 lines of code  estimated 797 years of effort (COCOMO model)  first commit in September, 2009  most recent commit about 13 hours ago
  • 7.
    Jan 1st 2015 RavenDB becomes self aware
  • 8.
    RavenDB 3.0  ~15team members  Some parts were started in 2011  18 months of work (another 2/3 remaining)  More than 600 issues  Awesome
  • 9.
    What did wedo?!  Voron  OWIN / Web API  Indexing  Operations  RavenFS  JVM client API  Spit & polish  New studio
  • 10.
  • 11.
    It’s not aboutthe UI  Yes, important  Yes, we have ~8 people on it now  We’re a database  http://issues.hibernatingrhinos.com  Features, not cosmetics
  • 12.
    Seriously, now…  Whatshould you be excited about?  It isn't the feature, it is the direction…
  • 13.
    Removing friction Indexing  Indexdeletes are async  Index ids  Small collection optimization  Fan out prevention Operations  No performance counters  Additional debug endpoints  Periodic backups bundle full/incremental & deletes  Explicit failover servers  Reduced # of assemblies  Server to server smuggling
  • 14.
  • 15.
    Raven File system Tailored persistence solution  In production since 2012  Replicated file system  Optimized change tracking  Very large files  Replacing attachments
  • 16.
    Spit & polish Preserving missing properties  Lazy async  Single pipeline (embedded / http)  Multiple database support for embedded  Everything on top of OWIN / Web API
  • 17.
    Where are wenow?  Stabilization  http://issues.hibernatingrhinos.com  Force new feature mode: On  No new features going in until release
  • 18.
    Hi, what aboutVoron?!  Internal only  Not important
  • 19.
    !!Not so fast! Voron is very important  Impl. details after lunch…  Implications of Voron are:  We own the entire stack  Tailored solutions
  • 20.
    Where are wegoing? Actual  Voron - distribution  Log shipping  Raft  Polyglot persistence solution  RavenFS is just the beginning  Event aggregations, the fallen feature Research  Project Corax  Project Tempus  Project Duco
  • 21.
    Hackaton  After hours Let us make something cool!  Full feature, from the disk to the UI
  • 22.