Is the Scality RING suitable for use as a private cloud archive?
Yes, the RING can be deployed on premises, on hardware you own, in your own data center. The RING can also be deployed in hybrid or public configurations, depending on your need.
What are the power and cooling usage stats for a RING system?
Deployed with SAS disks for primary storage of a cloud application, such as webmail or a photo sharing site, a small 100-node RING can deliver hundreds of thousands of objects per second, with a predictable latency below 40 milliseconds at a cost of less than $1/GB. When deployed with large SATA disks for long term storage such as cloud backup, Scality RING’s total cost of operation is only a few pennies per GB per month.
Although Scality RING is hardware agnostic and the costs depend on the hardware selected, they tend to go down proportionally as you scale up the storage usage of your system. The following diagram shows the dollar costs for Scality RING per GB of storage per month of total usable capacity in terabytes. When using 100 GB of storage per month, datacenter power and cooling costs are approximately three cents per GB. If you use 15 TB of storage per month, these costs go down to less than a penny per GB.
What kind of hardware can I use with RING?
When it comes to backend storage repositories, Scality RING is system-agnostic. It can utilize standard x86 servers, and can leverage any type of storage media, from SSD, to fast SAS disk, to denser and cheaper SATA disks. Scality RING can also benefit from the low latency of 10 Gigabit Ethernet and Infiniband networks. While traditional storage technologies are optimized for very fast serial operations, RING is a completely parallel design, which is ultimately a more effective solution when serving many users.
You can install Scality RING on a Linux (Centos/Redhat or Ubuntu) server, and manage the RING from any platform with a web browser.
Can I install and implement RING myself?
Instructions for installation are available on Scality wiki pages. Software updates are downloadable from the Scality repository website. These websites require a login and may be ACL-restricted. As with any carrier grade software or equipment, a minimum of a few days of design and integration services is recommended to make sure that the architecture and installation are reliable, sustainable and future proof.
Is RING suitable for database storage?
Scality RING software is not currently suitable for storage of traditional relational databases or virtual desktop infrastructures. However, it remains one of the best choices for both primary and long term storage of unstructured data. Unstructured data refers to information which does not fit well into a traditional relational database model. This includes files, whether they are text or documents of any sort, photos, videos, music, scan, data resulting from imaging or surveys, emails, html pages, logs and CDRs. In many cases, unstructured data actually does have an underlying structure but it is not easily handled by traditional relational databases.
Where are the Indexes in this system?
Indexes are user metadata that are kept on the RING for all stored objects. In the most recent releases, a special index daemon, sindexd, is included with the Scality RS2 connector. The sindexd daemon collects the metadata of the objects (key, file name, file size, modified time, owner, flags, version, and so on) stored on the RING or offloaded to second tier storage. Batched bucket indexing commands are sent to the sindexd indexing utility at intervals controlled by configuration settings. The ability of sindexd to process batched index requests and responses fulfills a basic requirement for extending RING support to big buckets (buckets that can contain millions of objects).