Comments on: Sharding destroys the goals of your relational database http://codemonkeyism.com/sharding-destroys-the-goals-of-your-relational-database/ Wed, 09 May 2012 22:39:34 +0000 hourly 1 http://wordpress.org/?v=3.3.2 By: 3.11.2011 I miss my best interview « luvSurfing http://codemonkeyism.com/sharding-destroys-the-goals-of-your-relational-database/comment-page-1/#comment-458508 3.11.2011 I miss my best interview « luvSurfing Mon, 21 Mar 2011 20:58:12 +0000 http://stephan.reposita.org/archives/2008/08/26/sharding-destroys-the-goals-of-your-relational-database/#comment-458508 [...] luvSurfboard Justine~(‾▿‾~) Sharding destroys the goals of your relational database:codemonkeyism.com/sharding-destr… #NoSQL 11 Mar luvSurfboard Justine~(‾▿‾~) Go Ahead: Next Generation Java [...] [...] luvSurfboard Justine~(‾▿‾~) Sharding destroys the goals of your relational database:codemonkeyism.com/sharding-destr… #NoSQL 11 Mar luvSurfboard Justine~(‾▿‾~) Go Ahead: Next Generation Java [...]

]]> By: stephan http://codemonkeyism.com/sharding-destroys-the-goals-of-your-relational-database/comment-page-1/#comment-318795 stephan Fri, 27 Aug 2010 12:50:53 +0000 http://stephan.reposita.org/archives/2008/08/26/sharding-destroys-the-goals-of-your-relational-database/#comment-318795 @Jens: I've meant code that you need to write, like handling joins in your application layer or write some general middleware. @Jens: I’ve meant code that you need to write, like handling joins in your application layer or write some general middleware.

]]> By: stephan http://codemonkeyism.com/sharding-destroys-the-goals-of-your-relational-database/comment-page-1/#comment-318794 stephan Fri, 27 Aug 2010 12:49:49 +0000 http://stephan.reposita.org/archives/2008/08/26/sharding-destroys-the-goals-of-your-relational-database/#comment-318794 @Sergio: "one of the most interesting patterns recently emerged." 100% agree @Sergio: “one of the most interesting patterns recently emerged.” 100% agree

]]> By: Jens Schauder http://codemonkeyism.com/sharding-destroys-the-goals-of-your-relational-database/comment-page-1/#comment-318790 Jens Schauder Fri, 27 Aug 2010 12:29:31 +0000 http://stephan.reposita.org/archives/2008/08/26/sharding-destroys-the-goals-of-your-relational-database/#comment-318790 you wrote about retrieval and searching, but what about aggregation? What solutions do you see for this? A little more specific than 'custom code' please :) you wrote about retrieval and searching, but what about aggregation?

What solutions do you see for this? A little more specific than ‘custom code’ please :)

]]> By: Sergio Bossa http://codemonkeyism.com/sharding-destroys-the-goals-of-your-relational-database/comment-page-1/#comment-318784 Sergio Bossa Fri, 27 Aug 2010 12:01:16 +0000 http://stephan.reposita.org/archives/2008/08/26/sharding-destroys-the-goals-of-your-relational-database/#comment-318784 Hi Stephan, pretty good post with interesting insights: on a related note, "splitting storage and searching" goes hand-in-hand with Command-Query Responsibility Segregation (see http://www.udidahan.com/2009/12/09/clarified-cqrs/ for an excellent post about it), one of the most interesting patterns recently emerged. Cheers, Sergio B. Hi Stephan,

pretty good post with interesting insights: on a related note, “splitting storage and searching” goes hand-in-hand with Command-Query Responsibility Segregation (see http://www.udidahan.com/2009/12/09/clarified-cqrs/ for an excellent post about it), one of the most interesting patterns recently emerged.

Cheers,

Sergio B.

]]> By: Dhananjay Nene http://codemonkeyism.com/sharding-destroys-the-goals-of-your-relational-database/comment-page-1/#comment-318783 Dhananjay Nene Fri, 27 Aug 2010 11:48:38 +0000 http://stephan.reposita.org/archives/2008/08/26/sharding-destroys-the-goals-of-your-relational-database/#comment-318783 Actually one context where this is a strength in a multitenancy configuration where the sharding is being implemented to isolate data storage for different companies to contain the risk of accidental data leakage. While it can constrain the host from running queries of type two and three the way you described, not being able to run them in the application then becomes a desirable feature. Is sharding the right way to describe this configuration. Probably not (at least since the intent is not the same as conventional sharding), even though the term does get used at times. Actually one context where this is a strength in a multitenancy configuration where the sharding is being implemented to isolate data storage for different companies to contain the risk of accidental data leakage.

While it can constrain the host from running queries of type two and three the way you described, not being able to run them in the application then becomes a desirable feature.

Is sharding the right way to describe this configuration. Probably not (at least since the intent is not the same as conventional sharding), even though the term does get used at times.

]]> By: stephan http://codemonkeyism.com/sharding-destroys-the-goals-of-your-relational-database/comment-page-1/#comment-166720 stephan Sun, 14 Sep 2008 18:35:14 +0000 http://stephan.reposita.org/archives/2008/08/26/sharding-destroys-the-goals-of-your-relational-database/#comment-166720 Thanks. Thanks.

]]> By: kyaw kyaw naing http://codemonkeyism.com/sharding-destroys-the-goals-of-your-relational-database/comment-page-1/#comment-166691 kyaw kyaw naing Sun, 14 Sep 2008 16:50:50 +0000 http://stephan.reposita.org/archives/2008/08/26/sharding-destroys-the-goals-of-your-relational-database/#comment-166691 "! Storage and search are two different things, if you split them, you gain flexibility" What a powerful concept. Your post is worth reading just for that quote. Thanks for other wonderful ideas, too. “! Storage and search are two different things, if you split them, you gain flexibility”

What a powerful concept.

Your post is worth reading just for that quote.

Thanks for other wonderful ideas, too.

]]> By: stephan http://codemonkeyism.com/sharding-destroys-the-goals-of-your-relational-database/comment-page-1/#comment-156861 stephan Thu, 28 Aug 2008 13:08:13 +0000 http://stephan.reposita.org/archives/2008/08/26/sharding-destroys-the-goals-of-your-relational-database/#comment-156861 @website: You're right, as I've said <i>"without custom code across"</i>. You always can write new code on top. @website: You’re right, as I’ve said “without custom code across”. You always can write new code on top.

]]> By: website designs http://codemonkeyism.com/sharding-destroys-the-goals-of-your-relational-database/comment-page-1/#comment-155544 website designs Tue, 26 Aug 2008 15:16:04 +0000 http://stephan.reposita.org/archives/2008/08/26/sharding-destroys-the-goals-of-your-relational-database/#comment-155544 Well, you just need some new layer to let you do summary queries across all your shards... Something like Map/Reduce. I can see LINQ going in this direction in the long term. Well, you just need some new layer to let you do summary queries across all your shards… Something like Map/Reduce. I can see LINQ going in this direction in the long term.

]]> By: todd http://codemonkeyism.com/sharding-destroys-the-goals-of-your-relational-database/comment-page-1/#comment-155538 todd Tue, 26 Aug 2008 14:54:56 +0000 http://stephan.reposita.org/archives/2008/08/26/sharding-destroys-the-goals-of-your-relational-database/#comment-155538 I think sharding and CAP come in because each shard can independently fail. So if 1 shard out of 10 fails your other 9 are still available, which increases the availability of your overall system at the expense of consistency. I think sharding and CAP come in because each shard can independently fail. So if 1 shard out of 10 fails your other 9 are still available, which increases the availability of your overall system at the expense of consistency.

]]> By: stephan http://codemonkeyism.com/sharding-destroys-the-goals-of-your-relational-database/comment-page-1/#comment-155436 stephan Tue, 26 Aug 2008 12:12:26 +0000 http://stephan.reposita.org/archives/2008/08/26/sharding-destroys-the-goals-of-your-relational-database/#comment-155436 @Arnon: I did know about CAP, but didn't see it in the case of shards. Interesting, thanks. @Arnon: I did know about CAP, but didn’t see it in the case of shards. Interesting, thanks.

]]> By: Arnon Rotem-Gal-Oz http://codemonkeyism.com/sharding-destroys-the-goals-of-your-relational-database/comment-page-1/#comment-155399 Arnon Rotem-Gal-Oz Tue, 26 Aug 2008 11:20:03 +0000 http://stephan.reposita.org/archives/2008/08/26/sharding-destroys-the-goals-of-your-relational-database/#comment-155399 The reason for this is CAP theorem Consistency, Availability and Partition tolerance you can only have two of the three. traditionally DBMSs take the first two. Shards mean taking the last two Arnon The reason for this is CAP theorem Consistency, Availability and Partition tolerance you can only have two of the three. traditionally DBMSs take the first two. Shards mean taking the last two

Arnon

]]>