Abstract: In a method and appliance for determining responsiveness of a service via a particular protocol, a device intermediary to a plurality of clients and a plurality of services determines response times from each of a plurality of services to respond to requests via a first type of protocol of a plurality of protocols. The device calculates an average response time for the first type of protocol from each of the response times of the plurality of services. The device establishes a predetermined threshold for which a response time of a service for the first type of protocol may deviate from the average response time. The device identifies a service as available responsive to determining that a deviation of the response time of the service from the average response falls within the predetermined threshold.

Abstract: The present invention provides systems and methods for maintaining site persistence in a hierarchical Global Server Load Balancing (GSLB) deployment. Via configuration of GSLB services locally and remotely on each of the GSLB appliances and LB appliances at a site, a site appliance identifies and associates requests from the GSLB with the site. Furthermore, the site appliance may receive a GSLB cookie with the client request and confirms the request is from the expected GSLB in the site hierarchy. When the load balancers receives a response from a server, the appliance may include the GSLB cookie with the response back to the client. The appliance may also include an LB cookie to identify the server selected by the LB. When the client sends another request, the request may include the GSLB and LB cookie. With this information, the GSLB and LB appliance may maintain site persistence for the client as well as server persistence at the site.

Abstract: The present solution provides centralized systems and methods for synchronizing a Global Server Load Balancing (GSLB) site hierarchy across a plurality of appliances in a multi-site deployment. With synchronization, differences in configurations across appliances may be determined and changes to configurations may automatically be distributed and applied on each appliance at each site to operate all the sites with a common single GSLB site hierarchy configuration. This reduces the challenging in configuration maintenance for the multi-site deployment.

Abstract: The present solution provides a centralized configuration of a Global Server Load Balancing (GSLB) site hierarchy that may be batched across a plurality of appliances in a multi-site deployment. A single GSLB site hierarchy configuration may be distributed and operated on each appliance at each site. This reduces the configuration maintained for the multi-site deployment. Furthermore, in this manner, each appliance across multiple sites has an understanding of the entire topology of the multi-site deployment. Responsive to this configuration and understanding, each appliance may perform and optimize operations according to the site topology. For example, the appliances may selectively determine which sites to establish connections with in order to share metrics. In another example, appliances may select which remote sites to monitor services according to the topology.

Abstract: The present invention provides improvements to the integration between a metric exchange protocol and a monitoring mechanism supporting load balancing by a Global Server Load Balancing (GSLB) appliance in a GSLB hierarchy of appliances, using a method for triggering a monitor on a remote service. The method includes receiving, by a first appliance, metrics of a second appliance via a metric exchange connection between the first and second appliances. The first appliance includes a GSLB service and identifies a service provided via the second appliance as a remote service for GSLB. The first appliance may identify via the metrics that a state of the service is up, and determine that the metric exchange connection is down. Responsive to the determination, the first appliance may trigger a monitor to monitor the status of the service. Further, the first appliance may obtain via the monitor a second status of the service.

Abstract: In a method and appliance for using a dynamic response time to determine responsiveness of one or more network services on a server, a monitor determines a response time for each of one or more network services. The monitor establishes an average response time based on the determined response time for each of the one or more network services. The monitor associates with the established average response time a predetermined threshold for which the response time of the one or more network services may deviate from the average response time and maintain an indication of responsiveness. The monitor monitors the response time of the one or more network services to one or more requests. The monitor determines that the response time of the one or more network services deviates from the average response time by the predetermined threshold. The monitor indicates that the one or more network services are unavailable.

Abstract: Systems and methods for load balancing services based on fewest connections by decreasing granularity of service selection as a number of fewest connections serviced by the services increases may include establishing, by an appliance, a set of identifiers corresponding to a number of connections serviced by a service, the set comprising a first plurality of identifiers each identifying a predetermined number of connections and a second plurality of identifiers each identifying a predetermined range of numbers of connections. The appliance assigns, to each service servicing connections, an identifier corresponding to the number of connections serviced by the service, at least one of the identifiers selected from the second plurality of identifiers. The appliance receives a request for a service, and forwards the request to a service assigned to the identifier corresponding to a fewest number of connections with at least one service assigned to the identifier.

Abstract: In a method and appliance for using a dynamic response time to determine responsiveness of one or more network services on a server, a monitor determines a response time for each of one or more network services. The monitor establishes an average response time based on the determined response time for each of the one or more network services. The monitor associates with the established average response time a predetermined threshold for which the response time of the one or more network services may deviate from the average response time and maintain an indication of responsiveness. The monitor monitors the response time of the one or more network services to one or more requests. The monitor determines that the response time of the one or more network services deviates from the average response time by the predetermined threshold. The monitor indicates that the one or more network services are unavailable.