Adoption of hyperconverged infrastructure (HCI) continues to soar, with a slew of vendors promising to eradicate storage complexity and to radically reduce IT costs. Unfortunately, many HCI solutions fail to keep promises made, especially as these architectures scale. Simplifying storage management lies at the heart of fulfilling these promises. The problem is that few vendors have taken steps beyond abstracting storage controller functionality into the hypervisor. The key to simplifying storage management is automation, and ultimately, artificial intelligence (AI) so that storage management is streamlined and resource utilization improves.
Hyperconverged solutions provider HiveIO’s Hive Fabric™ is built on the Linux-based Kernel-based Virtual Machine (KVM) hypervisor. It can be deployed on any x86-based hardware platform running the Ubuntu operating system, and it uses the GlusterFS open source and scalable file system management software. HiveIO describes the architecture as “zero layer,” meaning that it collapses expensive and cumbersome, legacy three-tier data center architectures into a singular fabric that can be easily scaled and adapted in tandem with dynamic business requirements. In following with this logic, the platform is sold as a single license.
In line with its hyperscale cloud-inspired roots, Hive Fabric supports multi-tenancy. It can create shared storage from individual server nodes as well as from memory (via a storage accelerator), and it applies a connection broker to broker virtual machines from a pool to users. For integration with third-party and customer applications, it supports a REST API and is able to provide customers and partners access to its Message Bus.
HiveIO’s Message Bus is Hive Fabric’s primary differentiator. The Message Bus uses HiveIO AI to streamline day-to-day infrastructure management, to enable IT managers to identify and react to forthcoming issues before they impact operations, and to provide greater visibility into how resources are being used. The Message Bus is a HiveIO construct that enables data to be shared between hosts, virtual machines and storage. The addition of AI enables the Message Bus to digest, format and analyze metadata pertaining to the state of the data center in real time.
Hive Fabric 7.0, released in August 2018, added the ability to support mixed application workloads on a singular infrastructure. It also introduced automatic creation and management of storage resources, as well as the Cluster Resource Scheduler, which intelligently migrates virtual machines between physical servers based on resource utilization without human interaction. These capabilities help to optimize workload performance while reducing the amount of physical hardware required to run these workloads (thus helping to improve the data center’s cost efficiency). They also help to streamline management and scalability.
HiveIO Hive Fabric 7.2
The newly-released Hive Fabric 7.2 user interface (UI) provides a centralized view into the orchestration, resource usage and impact on network input/output (I/O) performance of central processing unit (CPU), memory, storage capacity and virtual machine resources – as well as any associated alerts. Administrators can monitor performance inline, they can obtain a historical view into performance, and they can oversee how storage is pooled as well as virtual machine hosting. It was designed to make it faster and easier for storage managers to quickly dial into areas of the infrastructure that need attention.
In a world that is steadily becoming “AI-washed,” HiveIO’s Message Bus-driven approach is unique and stands to add value for the many enterprises that are struggling to get a handle on how to harness metadata to improve their resource utilization and IT service quality. Although these capabilities are currently largely reactive, HiveIO has alluded to plans to facilitate more proactive service management, to further empower infrastructure managers to keep pace with business modernization. HiveIO has a foothold in a number of large healthcare, education and service provider accounts that should yield telling use cases as its AI-driven management strategy continues to mature.
On the flip side, like many HCI vendors, HiveIO targets VDI and general virtualized workload hosting, and reduced complexity and greater cost-effectiveness when compared to virtualization incumbents (read: VMware). Against the backdrop of a highly competitive HCI vendor landscape, HiveIO faces the challenge of differentiating itself with customers as well as with partners such as hardware original equipment manufacturer (OEM) Lenovo and distributor Arrow Electronics. Its focus on carefully selecting its swimming lanes, rather than on going after multiple markets with multiple solutions, bodes well for its ability to bring differentiated, value-add storage management capabilities to the table.