International Business Machines Corporation

Alexander T Iannicelli

Marc Cochran

Martin L Schmatz

Malcolm Scott Allen-Ware

Mitch Gusat

Kevin Allen Hughes

Jason Garth Pearce

Jeffrey Bloom

Joseph Liu

Alan Ross

Akos Mate

James H H Chou

A virtual machine live migration is performed. A port mirroring rule for a source virtual machine in source computer is created by a computer system such that ingress traffic received at the source computer for the source virtual machine is delivered to both the source virtual machine and a target virtual machine on a target computer. The port mirroring rule is established prior to the virtual machine live migration from the source virtual machine to the target virtual machine. The source virtual machine on the source computer is migrated by the computer system to the target virtual machine on the target computer in which the port mirroring rule directs the ingress traffic to both the source virtual machine and the target virtual machine while the source virtual machine is migrated from the source computer to the target virtual machine on the target computer during the virtual machine live migration.

Virtual machine live migration with seamless network connectivity

Methods and apparatus are provided for managing data flows in a switch connected in a network. Such a method includes monitoring a set of data flows traversing the switch for compliance with a predetermined resource-usage policy, and, in response to detection of a non-compliant data flow, mirroring a set of data packets of that flow to send respective mirror packets to a mirror port of the switch. The method further comprises using the mirror packets sent to the mirror port to construct a non-compliance notification for the non-compliant flow, and sending the non-compliance notification into the network. The resource-usage policy can be defined such that the switch is operable to send a non-compliance notification before occurrence of congestion due to the non-compliant flow.

Flow management in networks

Preparing a plurality of computer nodes to boot in a multidimensional fabric network is provided. The method includes a fabric processor (FP) generating a plurality of DHCP discovery packets using a baseboard management controller (BMC) MAC address, and placing them into the multi-host switch. A dedicated connection directly connects the BMC and the FP. All ports of the multi-host switch broadcast DHCP discovery packets into the fabric network. The BMC, FP, and switch are all within the node. A designated exit node inside the fabric connects to a provisioning node not part of the fabric. The exit node relays DHCP traffic from the fabric. A location-based IP address uniquely identifies the nodes' physical location in the fabric. The IP address is calculated based on inventory records describing physical location information about the nodes. The FP calculates a host MAC address using its IP address and configures it onto the switch.

Preparing computer nodes to boot in a multidimensional torus fabric network

Growing community of inventors

Andover, MA, United States of America

Alexander T Iannicelli