Hitless restart mechanism for non-stop data-forwarding in the event of L3-mobility control-plane failure in a wireless switch

• US 20080019302A1
• Filed: 07/20/2006
• Published: 01/24/2008
• Est. Priority Date: 07/20/2006
• Status: Active Grant

• Current Assignee: Extreme Networks, Inc.
• Original Assignee: Symbol Technologies, Inc. (Zebra Technologies Corporation)
• Inventors: Nagarajan, Ramakrishnan, Borkar, Udayan
• Granted Patent: US 7,613,150 B2
• Time in Patent Office: Days
• US Class Current: 370/328
• CPC Class Codes: H04L 49/552 by ensuring the integrity of packets received through redundant connections
• CPC Class Codes: H04L 69/40 for recovering from a failure of a protocol instance or entity, e.g. service redundancy protocols, protocol state redundancy or protocol service redirection management of faults, events, alarms or notifications in data switching networks H04L41/06
• CPC Class Codes: H04W 84/12 WLAN [Wireless Local Area Networks]
• CPC Class Codes: H04W 88/12 Access point controller devices

Abstract

Techniques and technologies are provided for restarting a first wireless switch when the first wireless switch fails. For example, after the first wireless switch fails, the first wireless switch transmits a restart message to its peer wireless switches to begin re-establishing peering sessions with peer wireless switches. The first wireless switch resets its home wireless client database (HWCDb) with information from its data plane wireless client database (KWCDb), and transmits the updated information in its HWCDb to its peer wireless switches. The peer wireless switches store this information in their respective control plane wireless client databases (CPWCDbs), and initially mark the wireless client devices in the respective CPWCDbs as stale entries. Each of the wireless switches then exchanges information from their respective HWCDbs with the first wireless switch (and other peer wireless switches) so that each wireless switch eventually has a wireless client database (WCDb) comprising each of the wireless client devices in the mobility domain.

Claims

1. A method for restarting a first wireless switch when the first wireless switch fails, wherein the first wireless switch comprises a control plane comprising a control plane layer 3 (L3) mobility module and a control plane wireless client database (CPWCDb), and a data plane comprising a data-forwarder module and a data plane wireless client database (KWCDb), the method comprising:

restarting the control plane L3 mobility module of the first wireless switch, when the first wireless switch fails, and transmitting a restart message from the first wireless switch to the peer wireless switches of the first wireless switch to announce that the first wireless switch has restarted and to begin re-establishing peering sessions with peer wireless switches of the first wireless switch;

resetting a home wireless client database (HWCDb) of the first wireless switch with the information from the KWCDb, and transmitting a first message from the first wireless switch to peer wireless switches of the first wireless switch, wherein the first message comprises the information in the HWCDb of the first wireless switch;

receiving, at the peer wireless switches, information in the HWCDb of the first wireless switch, storing the information from the HWCDb of the first wireless switch in respective CPWCDbs of the peer wireless switches, and initially marking the wireless client devices in the respective CPWCDbs of the peer wireless switches as stale entries;

exchanging information from respective home wireless client databases (HWCDbs) of each peer wireless switch with the first wireless switch as a peering session is established between the first wireless switch and that peer wireless switch, wherein each respective HWCDb comprises the set of wireless client devices for which the particular peer wireless switch is the home wireless switch;

refreshing the CPWCDb of the first wireless switch and the KWCDb of the first wireless switch with updated information about each of the wireless client devices in the mobility domain such that the CPWCDb of the first wireless switch and the KWCDb of the first wireless switch are synchronized with information from the respective HWCDbs of the peer wireless switches, and transmitting the updated HWCDb of the first wireless switch to the peer wireless switches; and

refreshing stale entries in the respective CPWCDbs of the peer wireless switches with the updated information from the updated HWCDb of the first wireless switch such that the respective CPWCDb of the each peer wireless switch and the respective KWCDb of each peer wireless switch are synchronized with the updated HWCDb of the first wireless switch.

2. A method according to claim 1, wherein resetting a home wireless client database (HWCDb) of the first wireless switch with the information from the data plane wireless client database (KWCDb), comprises:

scanning the KWCDb at the first wireless switch, wherein the KWCDb comprises information for each of the wireless client devices for which the first wireless switch was either the home wireless switch and/or the current wireless switch before failing;

discarding the CPWCDb at the first wireless switch; and

updating the home wireless client database HWCDb of the first wireless switch with the information from the KWCDb of the first wireless switch.

3. A method according to claim 1, wherein transmitting a first message from the first wireless switch to peer wireless switches of the first wireless switch, wherein the first message comprises the information in the HWCDb of the first wireless switch, comprises:

transmitting a first message from the first wireless switch to peer wireless switches of the first wireless switch when the control plane L3 mobility module restarts, wherein the first message indicates that the first wireless switch has restarted and comprises the information in the HWCDb of the first wireless switch thereby allowing information to/from wireless client devices and to/from a wired network to continue to flow through the first wireless switch.

4. A method according to claim 1, wherein exchanging information from HWCDbs of each wireless switch with other wireless switches, comprises:

exchanging information from HWCDbs of each wireless switch with other wireless switches such that each wireless switch eventually has a wireless client database (WCDb) comprising each of the wireless client devices in the mobility domain, wherein each HWCDb comprises the set of wireless client devices for which the particular wireless switch is the home wireless switch.

5. A method according to claim 1, wherein refreshing the CPWCDd of the first wireless switch and the KWCDb of the first wireless switch with updated information about each of the wireless client devices in the mobility domain if the session is re-established within the restart period, comprises:

receiving, at the first wireless switch, the information about wireless client devices from the respective HWCDbs of the peer wireless switches, and updating the CPWCDd of the first wireless switch with the information about each of the wireless client devices; and

updating the KWCDb of the first wireless switch with the information about each of the wireless client devices in the mobility domain with information from the CPWCDd of the first wireless switch, determining ones of the stale entries in the KWCDd which are updated with information in the respective HWCDbs of the peer wireless switches and replacing the ones of the stale entries in the KWCDb with the updated information from the respective HWCDbs of the peer wireless switches, and removing any remaining stale entries from the KWCDb of the first wireless switch.

6. A method according to claim 1, wherein refreshing stale entries in the respective CPWCDbs of the peer wireless switches with the updated information from the updated HWCDb of the first wireless switch, comprises:

removing remaining stale entries from the respective CPWCDds and the respective KWCDbs of the peer wireless switches, wherein the remaining stale wireless client devices comprise ones of the stale entries originally learned by the peer wireless switches from the first wireless switch and not re-learned again from the first wireless switch after the peering sessions are re-established between the first wireless switch and the peer wireless switches.

7. A method according to claim 1, wherein the CPWCDd and the KWCDb function independently of each other so that failure of one does not affect the other.

8. A method according to claim 1, further comprising:

exchanging a hitless-restart capability message between each wireless switch in a mobility domain when the wireless switches originally begin a peering session, wherein the hitless-restart capability message indicates capability to perform hitless-restart,

wherein the hitless-restart capability message from the first wireless switch specifies a restart period of the first wireless switch,

wherein the restart period starts when the peer wireless device receives the restart message from the first wireless switch, and

wherein the first wireless switch and the peer wireless switches continue to retain the information about the wireless client devices during the restart period.

9. A method according to claim 8, further comprising:

determining, at the peer wireless switches, if a session with the first wireless switch is re-established within the restart period.

10. A method according to claim 1, further comprising:

forwarding information, from/to the data-forwarder module of the first wireless switch, to/from the wireless client devices while the first wireless switch is in the process of re-establishing peering sessions between the first peer wireless switch and the peer wireless switches of the first wireless switch.

11. A method according to claim 10, wherein the data-forwarder module of the first wireless switch continues forwarding information to/from the wireless client devices until the CPWCDd and the KWCDb are synchronized.

12. A method according to claim 1, wherein for each particular wireless client device, the information about the wireless client device comprises mobility related parameters to distinguish between new wireless client devices entering the mobility domain and existing wireless client devices roaming within the mobility domain.

13. A method according to claim 12, wherein the mobility related parameters for each particular wireless client device comprise:

a MAC address of the particular wireless client device,

an IP-address of the particular wireless client device,

an IP address of the home wireless switch (HS) for the particular wireless client device,

an IP address of the current wireless switch (CS) for the particular wireless client device, and

a VLAN identifier of the home wireless switch (HS) for the particular wireless client device.

14. A network of peer wireless switches in a first mobility domain, comprising:

a first wireless switch in the first mobility domain, wherein the first wireless switch comprises;

a first processor comprising a control plane layer 3 (L3) mobility module configured to restart after the first wireless switch fails;

a first transmitter configured to transmit a restart message to the peer wireless switches of the first wireless switch to announce that the first wireless switch has restarted and to begin re-establishing peering sessions with the peer wireless switches;

a first database comprising a first control plane wireless client database (CPWCDb), a first data plane wireless client database (KWCDb), and a first home wireless client database (HWCDb) configured to store a set of wireless client devices for which the first wireless switch is the home wireless switch, wherein the first HWCDb is reset with the information from the first KWCDb upon restart of the first wireless switch,

wherein the transmitter is configured to transmit a first message to the peer wireless switches, wherein the first message comprises the information in the first HWCDb after it is reset;

a first receiver;

wherein each peer wireless switch comprises;

a respective receiver configured to receive information in the HWCDb of the first wireless switch;

a respective control plane wireless client database (CPWCDb) configured to store the information from the HWCDb of the first wireless switch, and initially mark the wireless client devices in the respective CPWCDb as stale entries;

a respective processor;

a respective home wireless client database (HWCDb) configured to store a set of wireless client devices for which the particular peer wireless switch is the home wireless switch;

a respective transmitter configured to transmit information from the respective HWCDb to the first wireless switch and any other peer wireless switches when a peering session is established between the first wireless switch and that particular peer wireless switch;

wherein the respective receiver is configured to receive information from the respective HWCDbs of each the other peer wireless switches and the first wireless switch, wherein each respective HWCDb comprises the set of wireless client devices for which the particular peer wireless switch is the home wireless switch

a respective data plane wireless client database (KWCDb); and

wherein the first receiver is configured to receive information from the respective HWCDbs of each the other peer wireless switches, and wherein the first CPWCDb and the first KWCDb are configured to be refreshed with updated information about each of the wireless client devices in the mobility domain such that the first CPWCDb and the first KWCDb are synchronized with information from the respective HWCDbs of the peer wireless switches, and wherein the first transmitter is configured to transmit the updated HWCDb of the first wireless switch to the peer wireless switches; and

wherein each of the respective receivers is configured to receive the information in the first HWCDb after it is reset, and the information from the respective HWCDbs of each the other peer wireless switches;

wherein each of the respective processors is configured to refresh stale entries in its respective CPWCDb with the updated information from the first HWCDb such that the respective CPWCDb and the respective KWCDb are synchronized with the first HWCDb.

15. A network according to claim 14, wherein the first processor resets the first HWCDb with the information from the first KWCDb by scanning the first KWCDb, wherein the first KWCDb comprises information for each of the wireless client devices for which the first wireless switch was either the home wireless switch and/or the current wireless switch before failing;

discarding the first CPWCDb, and updating the first HWCDb with the information from the first KWCDb.

16. A network according to claim 14, wherein the first transmitter transmits the restart message when the control plane L3 mobility module restarts, wherein the restart message indicates that the first wireless switch has restarted and information to/from wireless client devices and to/from a wired network can continue to flow through the first wireless switch.

17. A network according to claim 14, wherein the first receiver is configured to receive the information about wireless client devices from the respective HWCDbs of the peer wireless switches, and wherein the first processor is configured to update the first CPWCDd with the information about each of the wireless client devices, and to update the first KWCDb with the information about each of the wireless client devices in the mobility domain with information from the first CPWCDd, and to determine ones of the stale entries in the first KWCDd which are updated with information in the respective HWCDbs of the peer wireless switches, and to replace the ones of the stale entries in the first KWCDb with the updated information from the respective HWCDbs of the peer wireless switches, and to remove any remaining stale entries from the first KWCDb.

18. A network according to claim 14, wherein each respective processor is configured to remove remaining stale entries from the respective CPWCDds and the respective KWCDbs, wherein the remaining stale wireless client devices comprise ones of the stale entries originally learned by the peer wireless switches from the first wireless switch and not re-learned again from the first wireless switch after the peering sessions are re-established between the first wireless switch and the peer wireless switches.

19. A network according to claim 14, wherein the first CPWCDd and the first KWCDb function independently of each other so that failure of one does not affect the other.

20. A network according to claim 14, wherein the first processor further comprises:

a data-forwarder module configured to forward/receive information to/from the wireless client devices while the first wireless switch is in the process of re-establishing peering sessions between the first peer wireless switch and the peer wireless switches of the first wireless switch, wherein the data-forwarder module continues forwarding information to/from the wireless client devices until the first CPWCDd and the first KWCDb are synchronized.