Today is a landmark day for Caringo, as we launch updates to our entire product suite. Our Swarm 10 Platform is the culmination of over a decade of market hardening and continuous innovation to satisfy customers who are driven by on-demand, distributed workflows. It leverages Swarm’s unique pure-object approach to deliver unrivaled performance at petabyte scale. Before we get into the details of today’s launch, let’s talk about what I mean by “pure-object.”
What is Swarm’s Pure-Object Architecture?
Caringo Swarm has always used a purpose-built storage design that efficiently utilizes standard disks without expensive caching layers, RAID controllers or layers of software. Swarm object storage software is a parallel architecture that boots from bare metal and runs completely in RAM. Swarm is native object storage, meaning each object’s metadata is encapsulated with that object’s data as a single entity. This means that no separate SQL databases or filesystems are required to track objects. With Swarm, objects are completely portable.
This approach extracts every bit of value from standard drives, servers and network infrastructures. It is at the core of our design and our key to delivering industry-leading S3 throughput with sustained PB-scale NFS read/write. Caringo SwarmNFS delivers a high-performance translation from file to object that uses patent-pending features in Swarm which eliminate the use of a spooler in the middle. Spoolers are expensive, put your data at risk, cause performance degradation, and get crushed under sustained ingest work loads.
Why the Updates?
Many content-driven organizations prefer to store data on premises to reduce security risks and copyright infringements. They need to cost-effectively scale to petabytes with distributed ingest and they need data to reside close to access points and applications. We are giving large organizations even more of what they already love about Caringo—performance, streamlined management and granular data insight.
At the same time, we are launching a new single-server appliance for smaller organizations giving them a much lower cost of entry to the most-scalable, on-prem, object-storage solution. This appliance will be valuable for a number of use cases, particularly for Media & Entertainment (M&E) organizations such as post-production houses, studios, broadcasters.
Highlights from our latest release include:
- Swarm 10 object storage has been optimized for dense, distributed environments, including an update to Elasticsearch 5. In a recent deployment, Swarm delivered an astounding 35 GB/s read and 12.5 GB/s write aggregate S3 throughput, the object storage industry’s fastest performance.
- New Swarm Single Server reduces entry-level hardware requirements by 75%. Swarm Single Server is a fully self-contained appliance that provides all the features of Swarm with 96 TB of raw storage that can be racked or fit under a desk.
- SwarmNFS 2.1 delivers parallel, petabyte-scale sustained streaming of NFS to object. In recent tests on standard hardware, a single instance of SwarmNFS sustained reads of 1.6 GB/s (3PB+ per month) with no caching or spooling. SwarmNFS also leverages Swarm’s parallel architecture so that multiple instances can be deployed as needed to further improve throughput.
- FileFly 3.0 now supports AWS, Azure and Google Cloud, providing file tiering from Windows and NetApp. In addition, a new, full-featured FileFly Community Edition is available that includes 25 TB of useable data transfer to any target.
Join me at 11am PT/1pm ET tomorrow for our webinar: Swarm 10—Storage for the On-Demand, Distributed World. Our VP of Marketing Adrian “AJ” Herrera and I will talk about what is new in Swarm 10 and detail how we have evolved our product line to set a new standard for on-premises object storage.
Do you need to securely store data and access it via NFS? In this webinar, John Bell and Michael Brame will discuss using NFS with object storage and explain how this has been handled by … More Details »
STFC uses Caringo Swarm object-based storage software to streamline data management and data access for long-term distribution. More Details »