Riak Hosting

Big data applications need power and flexibility. Boost the performance of your big data application with SoftLayer's robust, engineered bare metal servers optimized for Riak.

We developed these server configurations in strategic partnership with Basho, the creators of Riak. Each server, in every deployment, has been uniquely configured with computing and networking resources that give Riak its best performance. While other providers offer Riak on virtual servers, we use engineered bare metal servers with more storage, CPU speed, and RAM capacity ideal for Riak's highly scalable and high-volume read/write features. So you can be confident that you're deploying the highest-performing Riak solution—anywhere in the cloud.

Our unique design-and-build tool lets you create an environment suited for your unique application needs. And our automated platform will configure everything and deploy it in about two hours, complete with Riak or Riak Enterprise installed on each server, CentOS64 installed, and network configurations that connect each server into the cluster.

Easy-to-use design and configuration tool.

SoftLayer Riak solutions are available with a pay-as-you-go pricing model, billed by the month or the year, with no long-term contracts.

Our engineered servers are configured for an exceptional Riak experience. Each server comes with CentOS64 and your choice of Riak or Riak Enterprise already installed. Upgrading to Riak Enterprise provides additional features including multi-data center replication, monitoring, and 24×7 Basho support.

Choose the performance package that best fits your big data requirements. Packages are available with a pay-as-you-go pricing model, billed by the month or by the year with no long-term contracts. Click the Design Your Solution button and use our unique design-and-build tool to begin creating your Riak deployment.

Engineered Server Configurations

Riak Enterprise Pricing

Annual Subscription

Monthly Subscription

Our Riak solutions were developed in strategic partnership with Basho, so you get infrastructure that's been specifically designed to run Riak at its optimal levels. As with all of our solutions, these servers have full integration with the SoftLayer platform and feature pay-as-you-go pricing.

Bare Metal Servers for Better Performance

Your big data solution can now have the power of bare metal and the ease of the cloud. Our servers provide a higher level of performance than virtualized solutions are capable of delivering. And our bare-metal servers are all available in real-time with no downtime via our portal or API, so you don't have to sacrifice agility for performance.

Full Access and Control

You get the total access and control of your complete Riak solution, including total utilization of your hardware, total say in where your servers and replica sets are deployed, and the ability to further customize it to any specification or requirement.

Best Practices Solution

SoftLayer and Basho engineers collaborated to optimize the hardware and software of this solution to provide optimal Riak performance practices based on Basho's insight, expertise, and experience. The configurations themselves were developed to provide a range of options that reflect considerations that different scales of databases must account for. (As the size of a data set and number of indexes increase, additional RAM, CPU, and storage resources are required, but in particular proportions.)

Easy Design and Provisioning

Our new solution designer tool makes it easy to create and configure a customized Riak environment. And our unique capability in automated deployment, installation, and configuration means that a solution can be provisioned at the push of a button. And it also means that deployment is standardized, reducing the risk of human error.

Integrated Basho Support

Get additional features with Riak Enterprise including multi-data center replication, 24x7 phone and email support, insight from Basho's expert team, and a single bill for the entire solution.

Riak is an open-source, highly scalable distributed database based on the NoSQL database system.

Riak is considered a fault-tolerant database because it's distributed across many nodes and implemented without a master node. As such, it doesn't have a single point of failure. Primarily designed for distributed cloud data infrastructures, Riak provides superior performance in high-volume read and write applications and can be used to build cloud file systems.

The standard version of Riak is integrated with features such as MapReduce, multi-node clustering and a few others, whereas the enterprise version is enhanced with multi-datacenter replication, extended monitoring, 24x7 support, and professional services.

Simply, Riak is a flexible storage engine with a REST API, map/reduce capability, designed to be fault tolerant, distributed and ops friendly.

Popular use cases for Riak:

• Session Stores
  Riak was originally created to serve as a highly scalable session store. This is an ideal use case for Riak, which is a fundamentally a key value store. Since user/session IDs are usually stored in cookies or otherwise known at lookup time, Riak is able to serve these requests with predictably low latency. Riak's content-type agnosticism also imposes no restrictions on the value, so session data can be encoded in many ways and can evolve without administrative changes to schemas.
• Ad Serving
  A common use case for Riak is using it as a data store for serving ad content to many different web and mobile users with low latency. In Riak, advertising content - images or text - can be stored with unique keys, or with Riak-generated keys. Users will often make keys based on a campaign ID for easy retrieval.

Riak Features/Benefits

• Riak is built to operate in highly distributed environments, such as the cloud.
• Riak is masterless; it has no single point of failure.
• Riak is highly fault-tolerant due to its distributed nature. This means you can lose access to many Riak nodes, but never lose data.
• Riak makes data highly available for use in read and write-intensive web applications.
• Riak scales simply and intelligently, simply add a node to a cluster, with no downtime.
• Riak is incredibly stable; many see Riak as the "building blocks" of a cloud data infrastructure.
• Riak's elegantly simple design makes for a lower cost of operations.

What is Riak?

Riak is an open source, highly scalable, fault-tolerant distributed database. You can learn more by following the Riak Fast Track tutorial.

Is there a limit on how much data can be stored in Riak?

When using Bitcask, the default backend, there are two limiting factors when determining the amount of data a given Riak node can store. The first is the amount of memory on the node, and the second is the amount of disk space. Bitcask currently stores all of the keys, not the values, in memory to reference their location on disk, but the number of keys you can store will vary based on the size of the keys. There is a slight amount of overhead (40 bytes) to each key.

Basho offers a Cluster Capacity Planning document and a Bitcask Capacity Calculator to help plan for both factors.

Why are my Riak servers not public facing?

Riak servers run on our private network to ensure maximum security for the data stored there.

How many Vnodes should I order?

Order enough Vnodes to cover the amount of physical servers you want for your deployment. The number of physical servers cannot exceed the number of vnodes.

How does Riak control the growth of vector clocks?

Vector Clocks represent the history of an object using a list of identifiers, counters, and timestamps for each update. Riak regularly prunes the list based on four parameters which can be set per bucket. These parameters are:

• small_vclock
• big_vclock
• young_vclock
• old_vclock

The small_vclock and big_vclock parameters refer to the length of the vclock list. If the length of the list is smaller than small_vclock it will not be pruned. If the length is greater than big_vclock it will be pruned.

The young_vclock and old_vclock parameters refer to the timestamp per vclock entry. If the list length is between small_vclock and big_vclock the age of each entry is checked. If the entry is younger than young_vclock it is not pruned. If the entry is older than old_vclock then it is pruned. More aggressive vector clock pruning can be achieved by lowering the values of these four parameters.

In what scenarios might data become corrupted in Riak?

Data corruption in Riak is very rare and will generally only happen because of a failure of the storage backend, or a physical hardware failure. The underlying storage engines have their own methods handling durability (except Memory, that only lives in RAM.

Bitcask uses an append-only file format that is immutable once written. Incomplete entries at the end of the file will be ignored.

LevelDB has tunable durability and never write in place, instead appending to a log file or or merges existing files together to produce new ones. LevelDB recovery code uses checksums to detect this and will skip the incomplete records. You can also set verify_checksums to true forcing all data read from underlying storage to be verified against corresponding checksums.

Are there are any speed tests for Riak?

Take a look at Basho's open source benchmarking tool, Basho Bench. (Make sure to bench it on your target platform.)