Enabling Innovation in the SDDC Journey
Today we are announcing the open sourcing of an important pillar in the software defined data center (SDDC) story – Open Hardware Management Services (OHMS) available at https://github.com/vmware/OHMS/. Intended to manage both server and switch hardware in the context of SDDC, the software is the result of over 18 months of effort within VMware.
Based on an extensible architecture which includes both in-band and out-of-band management, the software provides server and switch management functionality. Servers can be managed through interfaces either in-band (with an operating system booted to run on the main CPU) or out-of-band (e.g. via the Baseboard Management Controller (BMC) on standard servers). An aggregator module smoothens out this dual interface so that you only use a single API for server management. Switches are managed using an out-of-band plugin, which in turn uses a switch-specific southbound interface (a switch API, switch CLI [Command Line Interface], etc.) to manage the switch.
OHMS includes an in-band plugin to communicate with ESXi running on the host and an out-of-band plugin that uses IPMI (Intelligent Platform Management Interface) to the BMC. The plugin framework allows for the addition of new plugins on both the in-band and out-of-band interfaces. For example, a new switch with a switch-specific config interface can be supported with a new plugin.
OHMS can be used to manage single-node or multi-node appliances (servers and switches), and even pods whose span can be across a data center (e.g. the full set of servers and switches across multiple racks). The software is scalable and also lends itself to multiple deployment models. For example, you can run the in-band and out-of-band agents together in a VM. Alternatively, you can run the out-of-band agent on a management switch. Clients of OHMS (including SDDC Management software) can use the OHMS north-bound API to manage hardware.
We are also happy to announce a contribution from Intel to the OHMS project, in the form of a Redfish plugin. This plugin can interact with emerging Redfish-compliant hardware. Intel is a key partner for VMware; the companies have worked together on several innovations over the years. We continue to collaborate in the fast growing area of SDDC/SDI (Software Defined Infrastructure). Learn more at our VMworld session about SDDC and Rack Scale Design.
Why are we open-sourcing hardware management services? This is a path we set out on last year, at VMworld 2015 to be specific. Our goal is to build a thriving, collaborative ecosystem of partners, customers, and other contributors who will accelerate ease of adoption of open standards-based hardware management. This will result in increased interoperability and the accelerated introduction of new hardware platforms and form factors. We would like to encourage partners, customers and other contributors to take this code, integrate it into their systems, add their own plugins and enhancements, all while conforming to the OHMS northbound API. VMware will continue to maintain this project and provide enhancements as well.
Again, here’s the link to the project: https://github.com/vmware/OHMS/. This project is the result of the hard work put in by several folks- kudos to the team.
Also, we have been busy building out the software to bring up, configure and manage (including life cycle operations) the software defined data center (SDDC). Combined with open hardware management innovations, expect to see SDDC become more pervasive. And do plan to attend VMworld Las Vegas (August 28 to September 1, 2016) if you would like to hear more.