Abstract: Example implementations relate to consensus protocols in a stretched network. According to an example, a distributed system includes continuously monitoring network performance and/or network latency among a cluster of a plurality of nodes in a distributed computer system. Leadership priority for each node is set based at least in part on the monitored network performance or network latency. Each node has a vote weight based at least in part on the leadership priority of the node. Each node's vote is biased by the node's vote weight. The node having a number of biased votes higher than a maximum possible number of votes biased by respective vote weights received by any other node in the cluster is selected as a leader node.

Abstract: A system for facilitating emulation in a distributed environment is provided. During operation, the system can obtain configuration of a physical edge rack, which can include one or more physical computing devices. The system can then present an emulated edge rack in a user interface. The emulated edge rack can be a digital twin of the physical edge rack. Based on a user input received via the user interface, the system can emulate a power-on event on the emulated edge rack. The emulated power-on event can emulate a power-on event for the physical edge rack. The system can then emulate the deployment of a component in the powered-on emulated edge rack based on configuration information received via the user interface. The deployed component can be configured based on the configuration information.

Abstract: Example implementations relate to consensus protocols in a stretched network. According to an example, a distributed system includes continuously monitoring network performance and/or network latency among a cluster of a plurality of nodes in a distributed computer system. Leadership priority for each node is set based at least in part on the monitored network performance or network latency. Each node has a vote weight based at least in part on the leadership priority of the node. Each node's vote is biased by the node's vote weight. The node having a number of biased votes higher than a maximum possible number of votes biased by respective vote weights received by any other node in the cluster is selected as a leader node.

Abstract: Example implementations relate to consensus protocols in a stretched network. According to an example, a distributed system includes continuously monitoring network performance and/or network latency among a cluster of a plurality of nodes in a distributed computer system. Leadership priority for each node is set based at least in part on the monitored network performance or network latency. Each node has a vote weight based at least in part on the leadership priority of the node. Each node's vote is biased by the node's vote weight. The node having a number of biased votes higher than a maximum possible number of votes biased by respective vote weights received by any other node in the cluster is selected as a leader node.

Abstract: A method of providing data stored in a network device to a subscriber, the method including providing a central database including data stored in a table, maintaining a sorted list for the table in the publisher, and generating a modification and a modification identifier for the table with the publisher. The method further including inserting the modification into the table in the central database based on the modification identifier, receiving a command from the subscribed for the table, and providing the table to the subscriber from the central database.

Abstract: A technique to manage a configuration database (CDB) for a network device is disclosed. Network devices may receive a configuration change request as a configuration change object. To process that request, a current configuration CLI set representative of the current CDB may be generated. The network device creates a shadow CDB initially corresponding to the current CDB and processes the change request against the shadow CDB. An updated configuration CLI set may then be generated from the updated shadow CDB. A differential CLI set indicating the difference between the first CLI set and the second CLI set may be generated to represent a set of CLI commands to transition from one CDB to the other (e.g., implement the request). Authorization of the user to execute the CLI commands of the differential CLI dataset may be verified. Upon verification, the current CDB may be replaced with the updated shadow CDB.

Abstract: A high-availability network device database synchronization technique for devices configured with multiple network controllers is disclosed. An HA database that contains information regarding a network state may not properly synchronize upon failure of a network component. For example, an HA switch typically has only two controllers, an active and a standby. If there is a loss of the active controller that causes a failover, changes in the network state may occur rapidly while the system is trying to recover (e.g., process the failover action). In part, because of the impact of the failover (e.g., failed communication paths) and rapidity of changes to network state while processing the failover, database changes may not be properly synchronized across all available database instances. Disclosed techniques provide reconciliation of database values using a mark and sweep technique on the “upside” of the failover and alter the “source of truth” for data value discrepancies.

Abstract: Methods of providing data to a user including receiving a page request from the user for a database to provide a sub-set of data from a dataset. The method further including sending a publish request from the database to a publisher, publishing the dataset from the publisher to the database, preparing the sub-set of data in the page request in the database, and sending the sub-set of data in the page request to the user.

Abstract: Example implementations relate to consensus protocols in a stretched network. According to an example, a distributed system includes continuously monitoring network performance and/or network latency among a cluster of a plurality of nodes in a distributed computer system. Leadership priority for each node is set based at least in part on the monitored network performance or network latency. Each node has a vote weight based at least in part on the leadership priority of the node. Each node's vote is biased by the node's vote weight. The node having a number of biased votes higher than a maximum possible number of votes biased by respective vote weights received by any other node in the cluster is selected as a leader node.

Abstract: One embodiment of the present invention provides a switch. During operation, the switch parses a first schema of the switch. The first schema indicates initialization information for one or more services of the switch expressed based on one or more tags. The switch then identifies a tag of the one or more tags in the first schema based on the parsing and identifies information corresponding to the tag from a profile of the switch. Subsequently, the switch generates a second schema from the first schema based on the identified information.

Abstract: A network switch may include a database, a network analysis data producer having a first application program interface, a network analysis data consumer having a second application program interface and a shared memory accessible by the first application program interface and the second application program interface. The network analysis data producer is to store network analysis data on a region of the shared memory and store an index to the region in the database for retrieval by the network analysis data consumer.

Abstract: A system for facilitating emulation in a distributed environment is provided. During operation, the system can obtain configuration of a physical edge rack, which can include one or more physical computing devices. The system can then present an emulated edge rack in a user interface. The emulated edge rack can be a digital twin of the physical edge rack. Based on a user input received via the user interface, the system can emulate a power-on event on the emulated edge rack. The emulated power-on event can emulate a power-on event for the physical edge rack. The system can then emulate the deployment of a component in the powered-on emulated edge rack based on configuration information received via the user interface. The deployed component can be configured based on the configuration information.

Abstract: Network device snapshots may capture the overall device state of a network device. Individual snapshots or groups of related snapshots (e.g., from different network devices obtained at a common time period) may be used to diagnose, troubleshoot, or correct anomalies or errors within a computer network. The “device state” of a network device may change over time and therefore obfuscate information desired for trouble shooting (e.g., diagnoses) of network errors (or degraded performance periods). Device state may include logical and physical device characteristics at a given instant in time. Network device snapshots may be stored locally on a network device or may be transmitted to external storage on-demand or periodically to accommodate possible limitations of resources on the network device. Network device snapshots may be “re-loaded” onto devices, for example in a lab or clean-room type environment, for comprehensive analysis. Different types of interfaces into network device snapshots are disclosed.

Abstract: One embodiment of the present invention provides a switch. During operation, the switch parses a first schema of the switch. The first schema indicates initialization information for one or more services of the switch expressed based on one or more tags. The switch then identifies a tag of the one or more tags in the first schema based on the parsing and identifies information corresponding to the tag from a profile of the switch. Subsequently, the switch generates a second schema from the first schema based on the identified information.

Abstract: A central database maintains values for client publishing devices or application processes executing locally with a database server process or remotely on a different device. Updates may be made by receiving, from a publishing application, an update to a statistic value in a database. The update may include: an indication to process the update as an accumulation update to the statistic value, an accumulation value to apply, and an indication of a key value pair to identify the statistic value to update. Upon receipt, a database server process may obtain access control as appropriate, obtain a current value for the update, apply the accumulation value to the current value, and store the result value as the new statistic value. Updates may be made by client publishers without client publishers being aware of a current value of the statistic. The statistic may represent a network wide operational parameter.

Abstract: Examples disclosed herein relate to a method comprising monitoring a first condition corresponding to a first parameter in a first database. The first database is configured to operate a first switch operating traffic on a network and the first database periodically synching with a second database. The method may also include monitoring a second parameter, corresponding to the first parameter, on the second database. The second database may be configured to operate a second switch on the network and the first switch and the second switch configured to operate traffic on the network. The method may also include determining that there is a deviation between the first and second that exceeds a threshold amount and transmitting an alert containing the deviation.

Abstract: Examples include re-connecting to a central database with data stored in a table. The table may include a total number of rows. Examples also include sending a message to the central database including a request for a first number of rows out of the total number of rows and receiving from the central database a second number of rows out of the total number of rows. Examples additionally include determining a non-existing row out of the second number of rows in a local table stored in a local cache and inserting the non-existing row into the local table.

Abstract: Network device snapshots may capture the overall device state of a network device. Individual snapshots or groups of related snapshots (e.g., from different network devices obtained at a common time period) may be used to diagnose, troubleshoot, or correct anomalies or errors within a computer network. The “device state” of a network device may change over time and therefore obfuscate information desired for trouble shooting (e.g., diagnoses) of network errors (or degraded performance periods). Device state may include logical and physical device characteristics at a given instant in time. Network device snapshots may be stored locally on a network device or may be transmitted to external storage on-demand or periodically to accommodate possible limitations of resources on the network device. Network device snapshots may be “re-loaded” onto devices, for example in a lab or clean-room type environment, for comprehensive analysis. Different types of interfaces into network device snapshots are disclosed.

Abstract: Networks that support business operations may be a complex combination of computers used by end-users, wired connections, wireless connections, and a multitude of infrastructure devices. Some of the infrastructure devices may be critical components in the operation of the network. Disclosed method and system provide for validation, deployment and rollback of configuration changes to network infrastructure components among other things. A computer device may include a network controller, memory storage for instructions and configuration database. A shadow database may be created to execute in parallel with the primary database service process, the shadow database instance comprising a shadow database control process and associated shadow database configuration information independently updatable from the configuration database information. Change validation may be performed using the shadow database without impact to the run-time configuration database.

Abstract: A network switch may include a database, a network analysis data producer having a first application program interface, a network analysis data consumer having a second application program interface and a shared memory accessible by the first application program interface and the second application program interface. The network analysis data producer is to store network analysis data on a region of the shared memory and store an index to the region in the database for retrieval by the network analysis data consumer.