In a previous post we briefly introduced Riak 2.0 data types. The addition of these distributed Data Types simplifies application development by automatically handling sibling resolution. This means developers can spend less time thinking about the complexities of vector clocks and sibling resolution and, instead, let Data Types support their applications’ data access patterns.
Understanding these data types requires a brief trip through history…
Riak 1.4 Counters
Riak 1.4 introduced counters as the first data types. Prior to 1.4 we’ve always said: “Your data is opaque to Riak,” — and it still can be — but with the addition of counters that is not longer the case. Riak knows what is stored in a counter key, and how to increment and decrement it through the counter API. It isn’t necessary to fetch, mutate, or put a counter. Instead you just incremented by 5 or decremented by 100. Vector Clocks, as discussed in the post entitled Clocks Are Bad, or, Welcome to the Wonderful World of Distributed Systems, as Riak knew how to merge concurrent writes there was never a sibling created.
Counters are very valuable, but you can not build many applications on just counters. Now, in Riak 2.0, we’ve added more data types. We believe that, with the addition of these data types you can model many applications’ data storage needs with greater simplicity, and never have to write sibling merge functions again.
What are CRDTs?
You may have heard a Basho presentation, or blog post, reference “CRDTs”. CRDT stands for (variously) Conflict-free Replicated Data Type, Convergent Replicated Data Type, Commutative Replicated Data Type, and others. The key, repeated, phrase is “Replicated Data Types”.
Replication is inherent in Riak. It is what the n-value defines. It is part of what lends to the availability and fault tolerance characteristics that Riak provides. Data Types are a common construct in computing. Sets, Bags, Lists, Registers, Maps, Counters…etc.
That leaves us to consider the “C”.
Conflict Free, or “Opaque No More”
Riak is an eventually consistent system. It leans, very much, towards the AP end of the CAP spectrum. (For more reading on the topic, the Practical Tradeoffs section of A Little Riak Book is particularly illuminating). This availability is achieved with mechanisms like sloppy quorum writes to fallback nodes. However, even without partitions and many nodes, interleaved or concurrent writes can lead to conflicts. Traditionally, Riak keeps all values and presents them to the user to resolve. The client application must have a deterministic way to resolve conflicts. It might be to pick the highest timestamp, or union all the values in a list, or something more complex. Whatever approach is chosen, it is ad-hoc, and created specifically for the data model and application at hand.
With Riak data types, there is still “conflict”. However, the resolution for that conflict is inherent and part of the data type’s design. The data types for Riak 2.0 converge automatically, at write and read time, on the server. If a client application can model its data using the data types provided, no sibling values will be seen and there is no longer a need to write ad-hoc, custom merge functions.
When modeling an applications data domain in a programming language, developers are familiar with composing state from a few primitive data types. Riak Data Types give the developer that power back and expressivity, and relieve them of the burden of design and testing deterministic merge functions. The key is that the data is no longer opaque to Riak. When the Data Types API is leveraged, Riak “knows” what type of thing is being stored and is able to perform the merge automatically.
When reading a Data Type from Riak, you will only ever see a single value. That value is still eventually consistent, but it will be as correct as it can be given the amount of entropy in the database. When the system is stable, all values will converge on a single, deterministic, correct value.
What Data Types Are Available?
Riak 2.0 includes the following Data Types:
- Counters: as in Riak 1.4
- Flags: enabled/disabled
- Sets: collections of binary values
- Registers: named Binary values with values also binary
- Maps: a collection of fields that supports the nesting of multiple Data Types
The conflict resolution, as discussed above, is intrinsic to the Data Type itself. This table provides greater detail.
|Data Type||Use Cases||Conflict Resolution Rule|
||Each actor keeps and independent count for increments and decrements. Upon merge, the pairwise maximum of any two actors will win (e.g. if one actor holds 172 and other holds 173, 173 will win upon merge)|
||Enable wins over disable|
||If an element is concurrent added and removed the add will win|
||The most chronologically recent value wins, based on timestamps|
||If a field is concurrently added, or updated and removed, the addd / update will win|
A new version of Riak, with new Data Types, allowing you to model your application in more expansive ways. Take these Data Types for a spin and be sure to let us know how you use them in your applications.
September features developer conferences, Chicago Erlang, and even an “unconference.” Take a look at where Basho will be around the U.S. this month.
Strangeloop (September 17-19 in St. Louis, MO): Strangeloop is a great opportunity to learn about emerging languages, concurrent and distributed systems, and new database technologies. Basho is attending, so tweet us @basho if you’re interested in meeting.
Analytics and Big Data Summit (September 18 in San Jose, CA at 3:05 p.m. PT): Produced by the Storage Networking Industry Association (SNIA), the Analytics and Big Data Summit brings together IT professionals to discuss how to leverage analytics, and big data applications and systems. Seema Jethani from Basho will be presenting on Optimizing Cloud Storage to Manage Big Data, which will explore different data types and storage solutions. Attendees will gain an understanding of the needs of big data storage and the current cloud storage options available to organizations.
2014 High Performance Computing for Wall Street (September 22 in New York, NY at 2:30 p.m. ET): The 11th annual HPC networking opportunity is focused on high put-through, low latency networks, data centers and lowering the costs of operations. Our director of technical marketing, Tyler Hannan, will be presenting a Code Writing Session – Architecting for Global Scale.
Chicago Erlang (September 22 in Chicago, IL): Chicago Erlang is a one-day event focused on real world applications of Erlang. At 10:40 a.m. CT, Basho’s Reid Draper will present on Building Fault Tolerant Teams at Basho during which he will explain how Basho coordinates the activities of more than 25 Erlang programmers to build Riak. Then, at 3:20 p.m. CT, Steve Vinoski from Basho will discuss Optimizing Native Code for Erlang.
REST Fest 2014 (September 25-27 in Greenville, SC): REST Fest is an “unconference” with the objective of bringing together people interested in REST, hypermedia APIs, web service APIs and related topics to share ideas, trade stories and show examples of current work. Sean Cribbs from Basho will be the opening keynote! His keynote, HTTP: The Good Parts, will explore interesting and powerful ways to enhance interaction and efficiency when developing applications. Sean will leverage his 10 years of experience as a developer to provide insight into HTTP features and how you can tap into them more declaratively.
Surge 2014 (September 24-26 in National Harbor, MD): We will be attending and sponsoring OmniTI’s scalability and performance conference, Surge. We’d love to meet and chat, so tweet us @basho if you’re attending.
Lastly, RICON 2014 is just one month away, October 28-29. Early bird prices are good through September 22. Register here.
February 5, 2014
At the recent meetup for the New York Linux Users Group (NYLUG), Basho Technical Evangelist, Tom Santero, presented “An Introduction to Basho’s Riak.” In this talk, Tom explains how Riak addresses the challenges of concurrent data storage at scale. He discusses the various design decisions, tradeoffs made, and theories at work within Riak. He also provides guidance as to how you might deploy Riak in production and why.
In addition to introducing the basics of Riak and its key/value data model, Tom presents some of the exciting features being introduced with Riak 2.0. Riak Data Types adds counters, sets, and maps to Riak – allowing for better conflict resolution. They enable developers to spend less time thinking about the complexities of vector clocks and sibling resolution and, instead, focusing on using familiar, distributed data types to support their applications’ data access patterns.
You can watch Tom’s full talk below:
For more information about Riak and how it differs from traditional databases, check out the whitepaper, “From Relational to Riak.”
To see where Basho will be presenting next, visit the Events Page.
January 29, 2014
On Fridays, Basho hosts a Hangout to discuss various topics related to Riak and distributed systems. While Basho evangelists and engineers lead these live Hangouts, they also bring in experts from various other companies, including Kyle Kingsbury (Fatcual), Jeremiah Peschka (Brent Ozar Unlimited), and Stuart Halloway (Datomic).
If you haven’t attended a Hangout, we have recorded them all and they are available on the Basho Technologies Youtube Channel. You can also watch each below.
Data Types and Search in Riak 2.0
Featuring Mark Phillips (Director of Community, Basho), Sean Cribbs (Engineer, Basho), Brett Hazen (Engineer, Basho), and Luke Bakken (Client Services Engineer, Basho)
Bucket Types and Configuration
Featuring Tom Santero (Technical Evangelist, Basho), Joe DeVivo (Engineer, Basho), and Jordan West (Engineer, Basho)
Riak 2.0: Security and Conflict Resolution
Featuring John Daily (Technical Evangelist, Basho), Andrew Thompson (Engineer, Basho), Justin Sheehy (CTO, Basho), and Kyle Kingsbury (Factual)
Fun with Java and C Clients
Featuring Seth Thomas (Technical Evangelist, Basho), Brett Hazen (Engineer, Basho), and Brian Roach (Engineer, Basho)
Property Based Testing
Featuring Tom Santero (Technical Evangelist, Basho) and Reid Draper (Engineer, Basho)
Datomic and Riak
Featuring Hector Castro (Technical Evangelist, Basho), Dmitri Zagidulin (Professional Services, Basho), and Stuart Halloway (Datomic)
Featuring John Daily (Technical Evangelist, Basho), David Rusek (Engineer, Basho), and Jeremiah Peschka (Brent Ozar Unlimited)
A Look Back
Featuring John Daily (Technical Evangelist, Basho), Hector Castro (Technical Evangelist, Basho), Andy Gross (Chief Architect, Basho), and Mark Phillips (Director of Community, Basho)
January 6, 2014
With the launch of the Technical Preview of Riak 2.0, we also announced the addition of strong consistency to Riak. This addition fundamentally changes how Riak can be used, since all previous versions classified Riak as an eventually consistent system.
With Riak 2.0, developers now have the flexibility to choose whether buckets should be highly available or strongly consistent, based on data requirements. Consistency preferences are defined on a per bucket type basis, in the same cluster.
At RICON West 2013, Basho senior engineer, Joseph Blomstedt, gave an updated version of his “Bringing Consistency to Riak” talk. The original talk (presented at RICON West 2012) discussed the challenges, motivations, and high-level plans of bringing consistency to Riak. This updated version presents the actual implementation that has since been built and how it will function in Riak 2.0. Both talks are available below.
To start testing the strong consistency feature, you can download the Technical Preview of Riak 2.0 here.
To watch all of the sessions from RICON West 2013, visit the Basho Technologies Youtube Channel.
December 17, 2013
In addition to Riak Data Types, there were a number of other presentations about Riak 2.0 features at RICON West. With the Technical Preview of Riak 2.0, we also announced a completely redesigned Riak Search.
Riak is a straight key/value data store and all objects are stored on disk as binaries. It is content agnostic – meaning you can store any type of data as the value in Riak. To improve the usability and functionality of Riak, we offer multiple querying options including Riak Search, Secondary Indexing, and MapReduce. Riak Search is a full-text search that allows Riak developers to index the contents of stored values. While Riak Search offers much needed functionality, it had its flaws.
In Riak 2.0, Riak Search received a complete overhaul. Riak Search 2.0 leverages the Apache Solr full-text document indexing engine directly. Riak users now get the power of Solr, with the availability and scalability of Riak. This upgrade also supports the Solr client-queries API, which enables integration with existing software solutions.
Eric Redmond is one of the Basho engineers who works on Riak Search. At RICON, he presented “Riak Search 2.0,” which walks through what’s new with Riak Search and why you’d want to use it. He also provides some impressive demos that show off the power of Solr and Riak. His full talk is below.
For more information on Riak Search 2.0, check out these resources on Github.
To watch all of the sessions from RICON West 2013, visit the Basho Technologies Youtube Channel.
December 12, 2013
At RICON West this year, we announced the Technical Preview of Riak 2.0. Before the full release (which will be available early next year), we are encouraging users to download the preview and start testing some of the exciting new features.
At RICON, we had many of the engineers who worked on these new features present their work. One feature that we’re particularly excited about is the addition of Riak Data Types. Riak 2.0 builds on eventually consistent counters (added with Riak 1.4) with the addition of maps and sets. These Riak Data Types simplify application development without sacrificing Riak’s availability and partition tolerance characteristics.
In “CRDTs: An Update (or Maybe Just a PUT),” Basho engineer, Sam Elliott, presents on the work being done with Riak Data Types. Sam and a few other engineers at Basho have been integrating cutting-edge research on data types (known as CRDTs), pioneered by INRIA, to create Riak Data Types. Sam talks about the latest developments on CRDTs and walks developers through how to use them in their own applications.
In addition to Sam’s talk, we also had a talk from Jeremy Ong on “CRDTs in Production.” His talk provides real world solutions to leveraging CRDT concepts for an industrial application via case study. He also offers some suggestions on how to tackle data operations that can’t always commute. You can watch his full talk below.
For more information about Riak Data Types, check out this overview on Github.
To watch all of the sessions from RICON West 2013, visit the Basho Technologies Youtube Channel.
November 18, 2013
This series of blog posts will discuss how Riak differs from traditional relational databases. For more information about any of the points discussed, download our technical overview, “From Relational to Riak.” The previous post in the series discussed High Availability and Cost of Scale.
In order to provide high availability, which is a cornerstone of Riak’s value proposition, the database stores several copies of each key/value pair.
This availability requirement leads to a fundamental tradeoff: in order to continue to serve requests in the presence of failure, we do not force all data in the cluster to stay in sync. Riak will allow writes and reads no matter how many servers (and their stored replicas) are offline or otherwise unreachable.
(Incidentally, this lack of strong coordination has another consequence beyond high availability: Riak is a very, very fast database.)
Riak does provide both active and passive self-healing mechanisms to minimize the window of time during which two servers may have different versions of data.
The concept of eventual consistency may seem unfamiliar, but if you’ve ever implemented a cache or used DNS, those are common examples of the idea. In a large enough system, it’s effectively the default state of all data.
However, with the forthcoming release of Riak 2.0, operators will be able to designate selected pieces of data to require coordination and maintain strong consistency over high availability. Writing such data will be slower and subject to failure if too many servers are unreachable, but the overall robust architecture of Riak will still provide a fast, highly available solution.
Riak stores data using a simple key/value model, which offers developers tremendous flexibility to define access models that suit their applications. It is also content-agnostic, so developers can store arbitrary data in any convenient format.
Instead of forcing application-specific data structures to be mapped into (and out of) a relational database, they can simply be serialized and dropped directly into Riak. For records that will be frequently updated, if some of the fields are immutable and some aren’t, we recommend keeping the immutable data in one key/value pair and the rest organized into a single or multiple objects based on update patterns.
Relational databases are ingrained habits for many of us, but moving beyond them can be liberating. Further information about data modeling, including sample configurations, are available on Use Cases section of the documentation.
One tradeoff with this simpler data model is that there is no SQL or SQL-like language with which to query the data.
To achieve optimal performance, it is advisable to take advantage of the flexibility of the key/value model to define simple retrieval patterns. In other words, determine the most useful queries and write the results of those queries as the data is being processed.
Because it is not always possible to know in advance what questions will need to be asked of your data, Riak offers added functionality on top of the key/value model. Tools such as Riak Search (a distributed, full-text search engine), Secondary Indexing (ability to tag objects with queryable metadata), and MapReduce (leveraged for aggregation tasks) are available to perform ad hoc queries as needed.
For many users, the tradeoffs of moving to Riak are worthwhile due to the overall benefits; however, it can be a bit of an adjustment. To see why others have chosen to switch to Riak from both relational systems and other NoSQL databases, check out our Users Page.
November 11, 2013
Last Friday, the Basho team held our inaugural Riak Community Hangout.
This 30 minute session is a development focused conversation with topics changing weekly. The Hangout is planned for most Fridays at 11am Pacific/2pm Eastern/7pm GMT, with the URL published shortly before it begins. All Hangouts will be archived and hosted on the Basho Technologies Youtube channel. You should follow @basho for all updates about future Hangouts.
Over the next few weeks, these Hangouts will focus on the new features planned for Riak 2.0.
The Hangout began with an overview of Riak Data Types, available with the 2.0 Technical Preview, and examined their implementation, use cases, and implementation considerations. Following this (at 18 minutes, 35 seconds), Brett Hazen provided an overview of Riak Search 2.0 (codenamed Yokozuna) and Luke Bakken queried a portion of the Twitter stream on a cluster running the newest Riak Search 2.0 code.
Upcoming sessions will focus on Riak/Riak CS internals, application building, data modeling, and community requested topics. We are also looking for community members to join in and highlight what you’re building with Riak and Riak CS developers
October 30, 2013
Today is the second day of RICON West, Basho’s distributed systems conference. If you weren’t able to make it out to San Francisco for the conference, tune in to our live stream to watch all of the great talks.
Day One of RICON West was an exciting one. Basho announced the availability of the Riak 2.0 Technical Preview (which Eric Redmond and Joseph Blomstedt described in more detail during their talks) and James Hughes of Seagate did the first live demo of the Kinetic Open Storage platform and Riak.
We also heard some amazing sessions from Pat Helland (Salesforce), Lindsey Kuper (Indiana University), Miles O’Connell (StackMob), Justin Shoffstall (Basho), Charlie Voiselle (Basho), Jeff Hodges (Twitter), Peter Bailis (UC Berkeley), Ryland Degnan (Netflix), and Derek Murray (Microsoft Research). We wrapped up Day One with Lightning Talks from across the industry and celebrated at Twenty Five Lusk (thanks for the drinks, Tower3 and Github!).
Today is another full day of amazing talks. Check out talks from Justin Sheehy (Basho), Diego Ongaro (Stanford University), Jeremy Ong (TBA), Jordan West (Basho), Susan Potter (Finsignia), Jason Brown (Netflix), Michael Bernstein (Code Climate), Sam Elliott (Basho), Andrew Thompson (Basho), Raja Selvaraj (The Weather Company), and Arvinda Gillella (The Weather Company). Additionally, Google Fellow, Jeff Dean, will be delivering the closing keynote on “The Tail at Scale: Achieving Rapid Response Times in Large Online Services.”