Startup vendor ioFabric Inc. on Tuesday previewed its flash-optimized Vicinity software designed to automate quality-of-service...
(QoS) management across distributed storage.
The vendor said a controlled release of Vicinity is planned for later this year, with general availability slated for 2016. The initial version profiles storage based on its performance and capacity characteristics. Profiling according to data protection requirements is expected in future releases.
IoFabric Vicinity QoS software identifies storage across cloud, flash, storage networks and even dynamic RAM. Then, the storage is profiled according to its location, performance characteristics and the applications it serves.
A micro-tiering engine intelligently places data on different types of storage media. Vicinity then provisions storage automatically based on an application's performance or capacity criteria.
Data-aware Vicinity runs as a physical or virtual appliance using industry-standard storage protocols, or installs as a software agent on Linux application servers in a converged storage model.
When installed on multiple servers, Vicinity agents combine to form a single data fabric that communicates via iSCSI. The QoS software discovers all available storage resources and assigns applications to the storage media that best meets required service levels.
For example, Vicinity will automatically move performance-starved applications to high-performance flash when it becomes available. It will raise alerts when it determines additional storage is needed to maintain application performance.
Vicinity discovers and profiles service levels for performance, capacity
According to ioFabric, its software allows storage admins to apply separate storage service policies to applications, either individually or as a group. It builds dynamic storage volumes that adjust in real time, as application workloads change. The software uses flash to buffer writes that are made to slow media in order to promote read-intensive workloads and to act as a flash write target for new workloads.
Automated QoS features in Vicinity include healing around drive failure, dynamic data routing and data migration to support rebuilds when a storage resource is unavailable. Administrators also could use a RESTful API or command-line interface to manually adjust settings in Vicinity's management dashboard.
Andrew Flint, vice president of marketing at ioFabric, said Vicinity is designed to anticipate which type of storage applications need and provision it before it's needed.
"We see storage as a key part of the infrastructure that you should be able to set and forget. It should be monitoring and managing itself," Flint said.
The ability to orchestrate storage across different platforms is a differentiating feature of ioFabric Vicinity, said George Crump, president of IT analyst firm Storage Switzerland LLC.
"You could almost call what they do software-defined QoS. If you look at the other vendors that have brought QoS features to market, most of them are focused on selling it with their own hardware," Crump said.
Vicinity's software agent can run as a virtual machine or on physical Linux and Windows servers, and supports VMware vSphere, Microsoft Hyper-V and open source kernel-based virtual machine (KVM) hypervisors.
Lamb said the first release of Vicinity is expected to support up to 64 physical nodes and 1 PB of active storage. He describes Vicinity as the storage equivalent to Google's driverless car technology.
"With cars, we've gone from the Model T Ford to the T-Bird to Google self-driving cars, which [replace] thousands of components with a simple dial to take me from A to B in X amount of time. We're looking at doing the same thing for storage," Lamb said.
IoFabric joins a crowded market for QoS software that includes software-defined storage vendors, legacy array makers and hyper-converged systems vendors. Lamb lists DataCore Software, EMC ScaleIO, Hedvig Inc. and Primary Data among his company's direct competitors.
Resource contention in SSD environments boosts need for QoS
Automation tools bring more effective QoS implementations
New storage technologies aren't so simple