Browse Prior Art Database

A Path Discovery Method in Overlay Virtual Network

IP.com Disclosure Number: IPCOM000241156D
Publication Date: 2015-Mar-31
Document File: 5 page(s) / 141K

Publishing Venue

The IP.com Prior Art Database

Abstract

In typical SDN environment, traffic from VM1 to VM2 will pass though some NVEs and waypoints. If some component fails, the traffic will be broken. The component can be the underlay network, NVE, or waypoint. There is no end to end diagnostic method. To find the packet path or debug the packet loss, this aritcle will give a innovative method, including 3 points: 1. A new trigger on NVE, when it’s enabled, NVE will process the overlay trace route packet. 2. innovated new format of ICMP TTL packet. 3. Replace the src IP with NVE underlay IP Add OTI(Overylay Traceroute Information) at the end of the packet.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 01 of 5

A Path Discovery Method in Overlay Virtual Network

At a high level, SDN overlay solution consists of attaching end stations to special overlay vSwitch and a group of controller to control these overlay vSwitch. The overlay vSwitch must have IP connectivity with each other. The physical infrastructure provides the required IP connectivity between the overlay vSwitch and can communicate with each other using IP. When an end station in a given virtual network transmits a packet to its peer, it is intercepted by the Overlay vSwitch, and Overlay vSwitch encapsulate the original packet sent by the source end station with an outer header that is constructed to transport the packet through the underlay. The packet then moves through the IP network, being forwarded/routed solely based on the outer header. It ultimately arrives at the destination vSwitch, which proceeds to strip the outer header and delivers the packet to the ultimate destination. To control the whether traffic could flow in the overlay Network. "Service Chaining" define an order of packet forward path. And user can insert their own appliances in the path to permit/deny the packet. These appliances are referred as "Waypoint". After "Service Chaining" is introduced, packets between 2 VMs will pass several "Waypoint". If the network administrator wants to find where packets are dropped during the path. Current this is no way to do it.

 Sending Multiple ICMP Echo packets with TTL=n, n=1,2,3,...

Finally, Sending ICMP Echo Reply when receiving the

ICMP Echo

Appliance

VM1/Host1

VM2/Host2

ICMP Echo with TTL=n Do TTL= Incoming TTL-1

Do TTL= Incoming TTL-1

Overlay vSwitch

Overlay vSwitch

Overlay vSwitch

Do TTL= Incoming TTL-1 Do TTL= Incoming TTL-1

Overlay Encap

Overlay Encap

Underlay Network

Trace Route Detecting Packet Flow

 Sending ICMP TTL Exceeded(type =11, code =0) when TTL=0

From the perspective of the detection initiation machine, the detecting method is the same as the traditional trace route, which utilizes the ICMP Echo or UDP with an unused destination port with increased IP header TTL (after here we call these packets as detecting packet) to trigger the network device in the whole path to send out ICMP Time Exceed packet or ICMP Port Unreachable packet (after here we call it as ICMP replay).

1



Page 02 of 5

The proposed method set each overlay vSwitch to trap all the detecting packets, which comes either from the underlay network port or overlay port. It will decrease the incoming packet's TTL by one, then if the TTL is equal to 0, it will send out an ICMP Time Exceed...