Big Compute for Big Data

Introducing the Joyent Manta Storage Service

Joyent Manta Storage Service is a highly scalable, distributed object storage service with integrated compute. Developers can store and process any amount of data at any time where a simple web API call replaces the need for spinning up instances. Joyent Manta Compute operates directly on objects with virtually zero data latency. It is a massively parallel, high performance compute environment including R, Python, node.js, Perl, Ruby, Java, C/C++, ffmpeg, grep, awk and others. Metering is by the second with zero provisioning, data movement or scheduling latency costs.


A Cloud Object Store

Store your objects across multiple data centers over a straightforward HTTPS REST API. Objects are stored and returned with their header metadata. The namespace is hierarchical and can be searched (yea!) Objects can be arbitrarily large and numerous.


Highly Durable

By default, Joyent Manta stores two copies of your data in two data centers. Each object can be stored with more or less redundancy, between one and six copies. On top of that, each copy is redundantly distributed across physical storage devices using erasure encoding to further survive unfortunate physical failures.


A Powerful Compute Service

The lightweight jobs interface works directly on the object store in multi-level map and reduce phases. Standard pipelining using stdio-like flows mean the most common open analytic tools work on Joyent Manta today.



Should you need to, it is also easy to add custom tools by including job-wide objects that can house applications and initialization scripts. You can even link to another user’s public tool objects to share scripts and applications.


Developer Friendly

If you’ve programmed for the web, you can use Joyent Manta as it is based on the Unix philosophy of offering small, well-crafted tools. Helper SDK support include: Node.js, Ruby, Java and Python, with more to come. Of course, the REST API is available to any language with HTTPS support.


Easy to Use Analysis

From real-time web page display decisions, to machine learning, to scientific and engineering analytics, Joyent Manta can help. The built-in compute service lets you effortlessly initiate hundreds of thousands of analytic tasks and gather them to create on-demand insights. Manta is a Platform as a Service for Big Data.


Compute on Data

Joyent Manta is designed with the novel idea of bringing compute to data instead of extracting data to do computation. Describing compute jobs is miniscule compared to large datasets. This means you now will have a durable system of record and a flexible analytics tool.


On-demand Jobs

Ad hoc inspection of a large dataset can be performed at or near interactive speeds. Compute resources are quickly run from pre-provisioned compute pools. Explicitly designed jobs can be refined over time and run repeatedly in a simple JSON pipeline. Tasks are billed out in seconds, so you don’t waste money on idle machines.


Manta Datasheet

Download the Manta Storage Service datasheet.

An integral part of the Joyent Cloud

The Joyent Manta Storage Service is just one of a family of services. Joyent Compute Services range from instances in our standard Persistent Compute Service (metered by the hour, month, or year) to our ephemeral Joyent Manta Compute Service (by the second). All are designed to seamlessly work with our Object Storage and Data Services.

Sign Up Now for Manta Storage Service Get Started

Includes 10GB Free Tier


Sign Up Now for Manta Storage Service Get Started

Includes 10GB Free Tier