About vNIOS Appliance for OpenStack
Infoblox vNIOS™ for OpenStack is a virtual appliance designed for deployment on Red Hat OpenStack Platform (RHOSP) with or without SR-IOV enabled and on Ubuntu-based OpenStack. Infoblox vNIOS for OpenStack enables you to deploy large, robust, manageable, and cost effective Infoblox Grids. For information about Infoblox Grids, refer to the Infoblox NIOS Documentation.
Infoblox NIOS provides core network services and a framework for integrating all the components of the modular Infoblox solution. NIOS provides integrated, secure, and easy-to-manage DNS (Domain Name System), DHCP (Dynamic Host Configuration Protocol) and IPAM (IP address management) services. NIOS also provides TFTP, HTTP, and FTP file transfer services. Infoblox vNIOS for OpenStack includes most features of NIOS, with some limitations described in Known Limitations.
You can configure most of the vNIOS appliances as standalone or HA (high availability) Grid Masters, Grid Master Candidates, and Grid members (or a reporting member). For supported vNIOS appliance models and their specifications, see vNIOS for OpenStack Virtual Appliance Models.
The performance of a vNIOS VM instance deployed on a OpenStack host is impacted by the huge page sizes configured on the instance and on the host, and also by the services enabled on the instance.
Although, huge page sizes configured on the host and the VM instance are independent of each other, the following behavior is observed:
When the host is configured to allocate huge pages to the instance, performance issues are reduced as it minimizes disk fragmentation, which otherwise has a significant impact on the performance. However, if the host is over provisioned with VM instances, performance of the VM will be impacted. To avoid it, refer to the best practices recommended by the Hypervisor.
When the host is not configured to allocate huge pages to the instance, performance issues can elevate due to increased possibility of disk fragmentation, which often leads to non-contiguous physical memory blocks being allocated to the VM. In some cases, this may also lead to memory allocation failures