Automatically selecting optimal transport protocol in a cloud computing environment

• US 9,075,645 B2
• Filed: 04/22/2013
• Issued: 07/07/2015
• Est. Priority Date: 01/23/2012
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Heninger, Ivan M., Hrischuk, Curtis, Jones, Zachary H., Quirk, Aaron J.
• Primary Examiner(s): Kawsar, Abdullah Al
• Assistant Examiner(s): TEETS, BRADLEY A
• Application Number:
  US13/867,585
• Publication Number:
  US 20130247041A1
• Time in Patent Office: 806 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 2009/45595 Network integration; Enabling network access in virtual machine instances
• CPC Class Codes: G06F 9/45533 Hypervisors; Virtual machine monitors
• CPC Class Codes: G06F 9/45558 Hypervisor-specific management and integration aspects

Abstract

A method, system and computer program product for selecting an optimal transport protocol in a cloud computing environment. An application creates a socket to communicate with a destination Virtual Machine (VM). The application requests a daemon agent to query a virtual machine manager to determine if the destination VM is on the same hypervisor as the application. The daemon agent receives a VMCI address (“context ID”) and an indication of the hypervisor associated with the destination VM from the virtual machine manager. The daemon agent enables the transport layer to route traffic over the VMCI protocol to the destination VM using its context ID in response to the destination VM being on the same hypervisor as the application. In this manner, the application is able to route traffic using an optimal transport protocol, including VMCI, depending on peer hypervisor association without requiring any code modifications to the application.

Claims

1. A method for selecting an optimal transport protocol in a cloud computing environment, the method comprising:

creating a socket to communicate with a destination virtual machine by an application;

requesting a daemon agent executed within a virtual machine hosting the application, by the application, to query a virtual machine manager to determine if said destination virtual machine is on a same hypervisor as said application, wherein the virtual machine manager executes independently from said same hypervisor;

receiving a context identification of said destination virtual machine and an indication of a hypervisor associated with said destination virtual machine from said virtual machine manager; and

enabling, by a processor, a transport layer to route traffic over a first transport protocol to said destination virtual machine using said context identification of said destination virtual machine in response to said destination virtual machine being on said same hypervisor as said application.

2. The method as recited in claim 1 further comprising:

enabling said transport layer to route traffic over a second transport protocol to said destination virtual machine using said context identification of said destination virtual machine in response to said destination virtual machine not being on said same hypervisor as said application.

3. The method as recited in claim 2, wherein said first transport protocol comprises a Virtual Machine Communication Interface (VMCI) protocol and a second transport protocol comprises a Transmission Control Protocol (TCP).

4. The method as recited in claim 1 further comprising:

deploying virtual machines that are endpoints in a communication to said same hypervisor in response to requirements of a service-level agreement being violated while using said virtual machines on separate hosts; and

enabling said transport layer to route traffic over said first transport protocol between said deployed virtual machines.

5. The method as recited in claim 1 further comprising:

opening a new connection over said socket;

sending a packet over said new connection;

inspecting said packet for its destination;

encapsulating said packet as a packet for said first transport protocol in response to said first transport protocol being established for said destination; and

sending said encapsulated packet over said socket.

6. The method as recited in claim 1 further comprising:

requesting all new traffic be sent over a socket for a second transport protocol unconditionally;

closing sockets for said first transport protocol once all pending packets have been sent over said sockets for said first transport protocol;

requesting an action involving a migration of one or more virtual machines between separate hosts;

notifying a daemon agent on a virtual machine residing on a hypervisor associated with said migrated one or more virtual machines that said action is completed; and

opening sockets for said first transport protocol for new packets with destinations on said hypervisor associated with said migrated one or more virtual machines.