Over a year ago, Qumulo released a new data-aware scale-out NAS solution designed to address what it considers a key weakness in scale-out storage – Increased data management complexity when adding nodes to the cluster. This means most scale-out NAS systems become less efficient and more expensive to operate as they grow. Administrators end up focusing more of their time managing storage instead of understanding how their data is used by users and departments, as well as its value to the organization.
A Quick Look Back
Qumulo Core is a data-aware, scale-out file and object storage software defined solution that runs on commodity hardware. It integrates the Qumulo Scalable File System (QSFS) with an object-based, flash-first hybrid architecture. QSFS captures information about data as it is being stored, and builds a “footprint” of that data. Qumulo leverages this information through a patented database built directly into the inodes of the file system, and designed to answer queries against that data.
However, not all data is created equal and not all data should be on flash storage, especially unstructured data. The Qumulo Core software is designed to support flash and hard disk drive tiers to properly balance cost and performance. But, maintaining this balance and making sure the right data is on the right media at the right time is the Achilles heel of most storage systems.
To ensure correct data placement, Qumulo Core provides real-time analytics integrated into the file system that provides visibility into data footprint, usage and performance. This enables the system to efficiently manage data transfers between the flash and hard drives in the Qumulo architecture. It also has a built-in interactive REST API browser making configuration and monitoring a simple process.
Additionally, the Qumulo Core 2.0 release provides mirroring and erasure coding for protection from media or node failure.
New Qumulo Core 2.5 Features
Qumulo‘s newest software release is Qumulo Core 2.5, which adds some new enhancements to an already robust solution.
- Snapshots, which provide extended data protection with API support allowing management by another process or system. They are pointer based using redirect-on-write and are read-only.
Erasure Coding improvements which deliver up to 80% efficiency by allowing wider stripes using a 8+2 scheme. The prior release was limited to a 4+2 scheme. Using the new 8+2 configuration however, will require a minimum of six nodes in a cluster. This allows data to be spread across a larger number of drives but without too many of them being located within a single node.
Throughput Analytics that show distribution of throughput load across the file system and provide greater insight into data usage and performance.
Intelligent caching of Metadata on SSD, providing 25x faster metadata operations.
While these are excellent enhancements, there are some limitations to take into consideration. Currently, there is no option available in the 2.5 release to convert an existing erasure coding scheme to the newer 8+2 scheme without a reinstall. While an in-place conversion option is on the Qumulo roadmap, there is no projected release date. Also the protection scheme must be selected at the time of installation.
Qumulo Core 2.5 now available on HPE Servers
In addition to the new software enhancements, the company is also making Qumulo Core 2.5 software available on HPE Apollo servers. This hardware option provides additional confidence for enterprise customers wanting next generation scale-out file and object storage on-premises or for private cloud workloads. It also shows the flexibility and portability of the Qumulo software-defined solution through its hardware independence.
Qumulo also continues to offer their QC series of hybrid flash-first commodity hardware appliances for on-premises and private cloud implementations. The QC series consists of the CQ24, CQ40, CQ104, CQ208, and CQ260 models with raw storage capacities ranging from 240TB to 2.6PB. Each unit has a high performance flash tier and a storage tier that uses the HGST 10TB hard drives. These units can be configured for mirroring or erasure coding data protection schemes.
Organizations drowning in the rising tide of unstructured data need a flexible, robust solution that not only enables them to store all this data securely but that also provides them the means to see how much capacity users, clients, and groups are using. It also provides insight into throughput showing how much data clients reading from or writing to any given data location. All of this enables organizations to see how their storage is being used.
The Qumulo Core 2.5 solution is one that organizations facing these data challenges should seriously consider.