More Proactive Infrastructure Operations and Cost-Effective, Granular Application Performance Acceleration
Graphics processing units (GPUs), software-defined data center architectures, and more intelligent infrastructure management are important technologies that all stand to help IT professionals meet modern business requirements. These include ever-faster levels of performance, increased infrastructure agility, and more time spent by IT staff on revenue producing initiatives as opposed to day-to-day management tasks. The problem is that the cost of new technologies like GPUs can outweigh the benefits if not implemented for maximum utilization and cost-efficiency. Additionally, many infrastructure management tools are not taking the required next steps toward becoming truly “intelligent” and predictive.
HiveIO Takes Steps Toward Composability and Proactive Management, and Adds SDN Support, with Hive Fabric 7.3
Intelligent hyperconverged fabric (HCF) provider HiveIO released version 7.3 of its infrastructure virtualization and datacenter management platform, Hive Fabric. Hive Fabric is built on the open-source KVM hypervisor and it uses the open-source GlusterFS file system. It can be deployed on any x86-based Ubuntu server. Hive Fabric’s core differentiator is its artificial intelligence (AI)-driven Message Bus. The Message Bus enables data sharing between server hosts, virtual machines (VMs), and storage infrastructure. The infusion of AI into the Message Bus helps IT to become nimbler, more proactive, and to support the business more strategically. It also provides better visibility into infrastructure resources. A previous Storage Switzerland briefing note contains a more in-depth overview of Hive Fabric.
Version 7.3 of Hive Fabric enhances HiveIO’s ability to simplify infrastructure operations, and to optimize IT infrastructure investments as well as performance, when it comes to instantiating virtual desktops, virtual servers, and software-defined storage. Communication between servers via the Message Bus is standardized in terms of format and in terms of types of information, such as the host ID, which is provided. This consistency makes this information easier for the Message Bus’s AI algorithms to consume. The Message Bus can also now dynamically allocate in-memory storage according to VM requirements. This configurability stands to improve the utilization of storage memory, which is one of the most expensive data center infrastructure resources.
The updated software platform also adds support for ATI, Intel and NVIDIA GPUs, for VM performance acceleration. According to HiveIO, up to 20% of its customers’ VMs require GPU acceleration to meet the levels of performance that their workloads require. The variety of GPU options is a differentiator and broadens the use cases that HiveIO can support (for example, ATI cards are less expensive than NVIDIA cards). HiveIO’s objective is to make the complex process of mapping GPU requirements and acceleration to VMs, as behind-the-scenes and simplified as possible. Users can turn acceleration on and off per VM with a single click.
With version 7.3, Hive Fabric also allows for multiple software-defined networks (SDNs) – whether physical or virtual – to be integrated into the HiveIO infrastructure. SDN implementations are flexible and can be configured per VM. The addition of SDN enables network traffic to be separated according to VM requirements, which helps IT professionals to meet performance requirements for applications and desktops. HiveIO also made SDN implementations easy to create, maintain and consume by enabling the bundling together of various networking interfaces.
With Hive Fabric 7.3, HiveIO also added a new capability called Hive Sense, bringing proactive support to the platform. Through Hive Sense, HiveIO proactively sends communications such as configurations, logs and metrics from customers back to HiveIO. This allows HiveIO to monitor a customer’s environment and also reduce the time it takes to resolve issues when they arise.
The adage “time is money” could not be truer for the modern IT professional, as businesses lean on IT strategically for competitive advantage. Many (arguably all) IT vendors today are touting some form of AI to vastly streamline the day-to-day life of the IT professional. But HiveIO’s infusion of intelligence into the Message Bus may facilitate a more user-driven model that can better handle distributed data environments.
HiveIO’s new Hive Fabric 7.3 capabilities are a notable step forward for HiveIO in delivering on its stated vision of providing flexible, easy-to-manage, low-cost and comprehensive virtualization. It also brings more proactive support capabilities to the table (both of which it discussed during a briefing with Storage Switzerland earlier in 2019). The composable-like capabilities for storage memory are of particular interest to Storage Switzerland when it comes to boosting agility and cost-efficiency. Additionally, HiveIO implements the new GPU and SDN support in a granular and flexible way that will better position IT to meet performance service level agreements (SLAs), in a cost-effective way.