Determining a transit appliance for data traffic to a software service

Abstract

Disclosed is a system and method for optimization of data transfer to a software service. In exemplary embodiments, a computer-implemented method for determining a transit appliance for data traffic to a software service through one or more interconnected networks comprising a plurality of network appliances, comprises determining performance metrics for each of the plurality of network appliances to at least one IP address associated with the software service, and selecting a transit appliance for data transfer to the IP address, the selected transit appliance based at least in part on the performance metrics.

Description

TECHNICAL FIELD

This disclosure relates generally to optimization of data transfer to a software service via a transit appliance.

BACKGROUND

The approaches described in this section could be pursued, but are not necessarily approaches that have previously been conceived or pursued. Therefore, unless otherwise indicated, it should not be assumed that any of the approaches described in this section qualify as prior art merely by virtue of their inclusion in this section.

Data centers may be used to provide computing infrastructure by employing a number of computing resources and associated components, such as telecommunication equipment, networking equipment, storage systems, backup power supplies, environmental controls, and so forth. A data center may provide a variety of services (e.g., web applications, email services, and search engine services) for a number of customers simultaneously. To provide these services, the computing infrastructure of the data center may run various software applications and store business and operational data. The computing resources distributed throughout the data center may be physical machines and/or virtual machines running on a physical host.

Computing resources of a data center may transmit and receive data packets via one or more interconnected networks, such as a Wide Area Network (WAN). Physical switches and routers can be distributed throughout the WAN and configured to connect various network segments and route the data packets within the network environment. It may be desirable to optimize or otherwise transform the data packets transmitted and received via the WAN. Routing of the data packets for optimization may be performed by configuring physical switches, routers, and/or other network appliances, to reroute the data packets to a data optimization virtual machine. However, involving reconfiguration of physical network components in data optimization may be costly and require complex coordination of various organizations and departments.

Additionally, an increasing number of computing resources and services are being hosted in the cloud. Infrastructure as a Service (IaaS) allows an organization to outsource the equipment used to support operations. As such, a request for a service may be first routed to a server associated with the service, with the server being housed in an IaaS center.

Software as a Service (SaaS) is also increasingly prevalent as it allows a user to access software services from any computing terminal. Access times for a user to access the SaaS may vary depending on the location from which a user is trying to access the software service. As the access time increases, the user may perceive performance and usability problems with the service. Furthermore, the software service hosted as SaaS may have its necessary computing equipment located in one or more physical locations, including IaaS locations. As such, a user request for a software service may first travel through one or more interconnected networks to one or more IaaS centers and then to the SaaS, which can be located anywhere in the world. Because the data may have to travel substantial geographic distances from each intermediate point, this increases the response time to the end user as well as the opportunities for packet loss.

While there are many optimization techniques that can be accomplished in a WAN, many of these optimization techniques for data transfer across a network require symmetric network components. For example, if data packets are encoded on the transmitting end before transmission through the network, they must be decoded on the receiving end. To optimize data transfer to a particular software service, it is desirable to decode the data as close to the requested software service as possible.

Therefore, a mechanism is needed to find an optimal transit appliance for a requested software service based on network performance characteristics, so that a user can access a software service with the most efficiency.

SUMMARY

This summary is provided to introduce a selection of concepts in a simplified form that are further described in the Detailed Description below. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used as an aid in determining the scope of the claimed subject matter.

In exemplary embodiments, a computer-implemented method for selecting a transit appliance for data traffic to a software service through a network comprising a plurality of network appliances, comprises: measuring one or more performance metrics of data traffic from at least one of the plurality of network appliances to an IP address associated with a software service, the IP address for the software service having been retrieved from a service directory; determining a derived performance metric to be advertised to the plurality of network appliances, the derived performance metric based at least in part on the one or more measured performance metrics; advertising the derived performance metric among one or more of the plurality of network appliances; updating an advertised metric table at one or more of the plurality of network appliances with the derived performance metric received from at least one of the plurality of network appliances; and selecting a transit appliance for data traffic to the IP address associated with the software service, the selection based at least in part on the advertised performance metrics. The performance metric may be based on at least one of network latency, data loss, and round trip time. The software service to be accessed may be hosted in a cloud-based environment. One or more of the plurality of network appliances may also be hosted in a cloud-based environment.

In further exemplary embodiments, the above method steps may be stored on a machine-readable medium comprising instructions, which when implemented by one or more processors perform the steps of the method. In yet further examples, subsystems or devices can be adapted to perform the recited steps. Other features, examples, and embodiments are described below.

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments are illustrated by way of example, and not by limitation in the figures of the accompanying drawings, in which like references indicate similar elements and in which:

FIG. 1A depicts an exemplary system environment for determining a transit appliance for transfer of data traffic to and from a software service.

FIG. 1B depicts an exemplary system environment for determining a transit appliance for transfer of data traffic to and from a software service.

FIG. 2 illustrates an exemplary service directory from the portal.

FIG. 3 illustrates an exemplary measured metric table at an appliance.

FIG. 4 illustrates an exemplary advertised metric table at an appliance.

FIG. 5 is a process flow diagram illustrating an exemplary method for the determination of a transit appliance to a software service.

FIG. 6 shows an exemplary system environment suitable for implementing methods for optimization of data across one or more interconnected networks.

FIG. 7A is a process flow diagram illustrating an exemplary method for the transmission of data packets for a software service via a first appliance.

FIG. 7B is a process flow diagram illustrating an exemplary method for the transmission of data packets for a software service via a transit appliance.

FIG. 8 is a screenshot of an exemplary GUI for a user to select optimization of data traffic to and from particular software services.

FIG. 9 shows an exemplary global network of appliances in an overlay network.

DETAILED DESCRIPTION

The following detailed description includes references to the accompanying drawings, which form a part of the detailed description. The drawings show illustrations, in accordance with exemplary embodiments. These exemplary embodiments, which are also referred to herein as “examples,” are described in enough detail to enable those skilled in the art to practice the present subject matter. The embodiments can be combined, other embodiments can be utilized, or structural, logical, and electrical changes can be made without departing from the scope of what is claimed. The following detailed description is therefore not to be taken in a limiting sense, and the scope is defined by the appended claims and their equivalents. In this document, the terms “a” and “an” are used, as is common in patent documents, to include one or more than one. In this document, the term “or” is used to refer to a nonexclusive “or,” such that“A or B” includes “A but not B,” “B but not A,” and “A and B,” unless otherwise indicated.

The embodiments disclosed herein may be implemented using a variety of technologies. For example, the methods described herein may be implemented in software executing on a computer system or in hardware utilizing either a combination of microprocessors or other specially designed application-specific integrated circuits (ASICs), programmable logic devices, or various combinations thereof. In particular, the methods described herein may be implemented by a series of computer-executable instructions residing on a storage medium, such as a disk drive, or computer-readable medium.

The embodiments described herein relate to computer-implemented methods for optimization of data transfer to a software service via a transit appliance.

FIG. 1A depicts an exemplary system environment for determining a transit appliance for transfer of data traffic to and from a software service across one or more interconnected networks 120, such as the Internet, or any other wide area network. In the exemplary embodiment, a user at computer 102a may access a software service, such as software service 110A, software service 1106, or software service 110N. While three software services are depicted here, there can be any number of software services. Data packets from the user at computer 102a may be transmitted via appliance 104a, which may also be referred to herein as the first appliance or ingress appliance for the request. Data packets from the user at computer 102b may be transmitted via appliance 104b, which is the first appliance or ingress appliance for that request.

The data packets from the user are then transmitted across the one or more interconnected networks 120, where there may be one or more peer appliances at different locations. In various embodiments as discussed herein, each of these peer appliances are in communication with each other, and form an overlay network that optimizes communication between the appliances. For example, the appliances may transfer data packets within the overlay network using one or more data transfer optimization techniques, such as compression/decompression, deduplication, TCP acceleration, performance enhancing proxy, packet reconstruction, error correction, or any other technique for optimizing data transfer between network appliances or devices.

Embodiments of the present disclosure provide for the selection of a transit appliance (also referred to herein as a second appliance or egress appliance), for each software service. The selected transit appliance (also referred to herein as the optimal transit appliance) may be the appliance which has the best network performance metrics for providing access to the requested software service, or component of the requested software service. In the exemplary embodiment depicted in FIG. 1A, the optimal transit appliance for software service 110A is appliance 106a, the optimal transit appliance for software service 110B is appliance 106b, and the optimal transit appliance for software service 110N is appliance 106n. As discussed further herein, appliances 106a, 106b, and 106n can be geographically located anywhere in the world.

FIG. 1B depicts an exemplary system environment for determining a transit appliance for transfer of data to and from a software service across the one or more interconnected networks. In the exemplary embodiment, an end user accesses a software service, such as software service 110A, through computer 102. Computer 102 may be a desktop computer, laptop computer, handheld computing device, server, or any other type of computing device. While a single computer is depicted here, computer 102 may also be a cluster of computing devices.

The request for software service 110A from computer 102 is transmitted via appliance 104, which is in communication with computer 102 through a network 108. The network 108 may include one or more of the following: WAN, the Internet, Metropolitan Area Network (MAN), Backbone network, Storage Area Network (SAN), Advanced Intelligent Network (AIN), Local Area Network (LAN), Personal Area Network (PAN), and so forth.

Appliance 104 can be any type of hardware device, or software operational on a computing device. Appliance 104 may be located at the same geographical location as computer 102, or may be located in a remote location. Appliance 104 may be in communication with other appliances across the network, such as appliances 106, 116, and 112, regardless of geographical location of the appliances. While appliance 104 is in communication with three other appliances in the exemplary embodiment depicted in the figure, there may be any number of appliances in the system. The appliances together may form an overlay network over the one or more interconnected networks between computer 102 and software service 110A.

Each of the appliances in the system may further be in communication with a portal 114. Portal 114 comprises a database with a service directory for the various software services, the IP addresses/subnets associated with each software service, and one or more test methods for determining network performance characteristics for each appliance in relation to the IP addresses/subnets associated with each software service. Portal 114 is also discussed in further detail below with respect to later figures. Each appliance in the overlay network is in communication with the portal 114 and retrieves a copy of the service directory. In various embodiments, the service directory is stored locally at each appliance, and the local copy at each appliance is updated on a fixed periodic schedule, upon a change in the service directory, upon the direction of a network administrator, or other triggering event. Exemplary changes in the service directory include the addition of a new software service, deletion of a software service, a change in an associated IP address/subnet, or a change in a test method.

Software service 110A may have an exemplary IP subnet of a.b.c.d/24. An appliance may query an IP address from the IP subnet using the information from the portal service directory to determine network performance characteristic(s) for the transmission of data between that appliance and software service 110A. The performance metric comprises information such as latency, round trip time, data loss, or any other network performance characteristic. The appliance then stores the measured performance metric(s) for the IP address or subnet in a local network performance characteristics table or database, referred to herein as a measured metric table. The measured metric table is discussed in further detail below in connection with FIG. 3.

FIG. 2 shows an exemplary service directory 200 from the portal 114. In various embodiments, the service directory 200 provides a listing of the various software services that are available for optimized access through the overlay network of appliances. There can be any number of software services in the service directory, such as software service 110A, 1106, and 110N. Each software service can have one or more IP addresses or IP subnet associated with it. The IP addresses may be in IPv4, IPv6, or other network addressing systems. While the term “IP address” has been used throughout this disclosure, a person of ordinary skill in the art would understand that any other network addressing system besides IP is also within the scope of this disclosure.

The service directory 200 can provide a listing of each IP address or subnets associated with the software service, one or more test IP addresses, and one or more test methods for the IP addresses. In various embodiments, additional data associated with each software service is also stored in the service directory 200, as understood by a person of ordinary skill in the art. The service directory 200 can be updated on a fixed periodic schedule, upon certain trigger events, or as directed by a network administrator.

In the exemplary service directory 200 of FIG. 2, software service 110A has two exemplary IP subnets, 152.3.4.0/24 and 97.5.6.0/24. The subnets may be in different geographical locations. Each IP subnet has one or more test methods associated with it. The test method can be ping IP, http-ping IP, tcp-ping IP, or any other test method as understood by a person of ordinary skill in the art. Each test method denotes the mechanism whereby the appliance queries an IP address associated with the software service to determine network performance characteristic(s) for the transmission of data from that appliance to the software service. As understood by a person of ordinary skill in the art, the subnet may contain many IP addresses. The test method may sample one or more of the included IP addresses, as testing every IP address in the subnet may introduce more traffic and overhead for very little additional information. Also, one service may have different optimal transit nodes for different parts of the service.

FIG. 3 shows an exemplary measured metric table 300 stored at each appliance for collecting measured metrics from that appliance to each software service. For each software service, the IP address or subnet associated with that software service is noted, along with the test method(s) used. The listing of software services, IP address/subnet, and test method(s) may be retrieved by each appliance from the service directory 200 in portal 114. In various embodiments, the measured metric table 300 is updated on a periodic fixed schedule, upon direction by a network administrator, or upon another triggering event, such as a change or addition of a subnet. Upon receipt of new information or a new service directory 200 from the portal 114, the information may be merged into the measured metric table 300 such that previous information from the measured metric table is maintained, if still applicable. Additionally, while the measured metric table and all other tables are described herein as “tables”, the data can also be represented using other data structures, as understood by a person of ordinary skill in the art.

In exemplary embodiments, an appliance queries one or more IP addresses associated with each software service in the table using the one or more test methods and measures one or more network performance characteristics. These characteristics may be stored in the measured metric table 300 as the measured metric(s). A derived metric related to the measured metric(s) is also stored in the measured metric table 300. The derived metric is a calculated or selected metric value that may be advertised, along with the corresponding tested IP address or subnet, with other peer appliances in the overlay network.

In the exemplary embodiment of FIG. 3, software service 110A has two associated IP subnets. For the IP subnet 152.3.4.0/24, the appliance queries the IP address 152.3.4.5 using a ping test method and measures a network performance metric of 70 milliseconds. For the IP subnet 97.5.6.0/24, the appliance queries the IP address 97.5.6.50 using an http-ping test method and measures a metric of 80 milliseconds, and also queries the IP address 97.5.6.51 using a tcp-ping test method and measures a metric of 70 milliseconds. In various embodiments, each measured metric may be stored in the measured metric table 300 fora fixed period of time, upon expiry of which it may need to be measured again. Additionally, the measured metric table 300 may keep a rolling average or other statistical aggregation for each measured metric instead of only the latest measured value(s). The statistical aggregation may be reflected in the measured metric(s), derived metric, or an additional field in the measured metric table 300.

From the various measured metrics, a derived metric may be calculated or selected for each tested IP address or subnet. The derived metric may be an average, mean, median, or any other statistical or calculated value from the one or more measured metrics. In the exemplary embodiment of FIG. 3, the derived metric for the IP subnet 97.5.6.0/24 is based on a scaled average of the two measured metrics from the http-ping and tcp-ping test methods. The derived metric may be updated on a periodic fixed schedule, as directed by a network administrator, or upon a triggering event, such as a change in a measured metric value.

In exemplary embodiments, the derived metric is then advertised by an appliance with the other appliances in the overlay network. For example, in the exemplary system environment of FIG. 1B, appliance 104 advertises one or more of its derived metric(s) for the IP address/subnet associated with software service 110A, with the peer appliances 106, 116, and 112. Similarly, one or more of the other appliances 106, 116, and 112 may also advertise one or more of their derived metric(s) for the IP address/subnet associated with software service 110A with all other peer appliances in the network. In various embodiments, an appliance may advertise all of the derived metrics for a particular software service, or only advertise a derived metric that is closest to a specified value, or a derived metric representative of the most ideal network characteristics, such as the lowest value or highest value. The derived metrics may be advertised to the other peer appliances on a periodic schedule, as directed by a network administrator, or upon a triggering event, such as a change in a derived metric value. Furthermore, if the derived metric is below or above a certain threshold, it may not be advertised with the other peer appliances.

FIG. 4 shows an exemplary advertised metric table 400 for collecting advertised metrics from the appliances in the overlay network. While FIGS. 2-4 have been described herein as a “table,” the data can be represented by other data structures as well, as understood by a person of ordinary skill in the art.

The advertised metric table 400 shows that for exemplary subnet a.b.c.d/24, peer appliance 104 has advertised a performance metric of 5, peer appliance 106 has an advertised performance metric of 20, peer appliance 116 has an advertised performance metric of 10, and peer appliance 112 has an advertised performance metric of 7.5. In various embodiments, a transit appliance for each IP subnet is selected based on the peer appliance with the lowest value advertised metric, the highest value advertised metric, or the advertised metric that is closest to a specified value. The specified value can be any value determined by a network administrator. In the exemplary embodiment of FIG. 4, the selected performance metric for subnet a.b.c.d/24 is the lowest value of 5, which corresponds to appliance 104. As such, appliance 104 is the optimal transit appliance to route data traffic through for the subnet a.b.c.d/24. In the exemplary embodiment of FIG. 4, the selected metric is noted by a box around the number. In other embodiments, the selected metric can be noted by any other means. Additionally, the advertised metric table 400 may optionally comprise one or more additional columns to note the peer appliance with the selected metric for the IP subnet as the optimal transit appliance, or to store any other information.

For exemplary subnet e.f.g.h/20, peer appliance 104 has an advertised metric of 15, peer appliance 106 has an advertised metric of 20, peer appliance 116 has an advertised metric of 10, and peer appliance 112 has an advertised metric of 8. If the selected performance metric is taken as represented by the lowest value, then peer appliance 112 is the selected transit appliance for the subnet e.f.g.h/20.

Advertised metric table 400 may be stored locally at each appliance, or stored in another central location that is accessible by all of the peer appliances, or stored and shared between appliances in other ways. In various embodiments, the table is updated on a periodic schedule, upon direction by a network administrator, or upon another triggering event, such as a change or addition of a subnet, peer appliance, or updated advertised metric. In various embodiments, each peer appliance's advertised metric may be stored in the advertised metric table 400 for a fixed period of time, upon expiry of which it may need to be updated. Additionally, the advertised metric table 400 may keep a rolling average or other statistical aggregation for each advertised metric instead of only the latest advertised values.

Now referring to FIG. 5, a flowchart 500 showing an exemplary method for the determination of a transit appliance to a software service is presented. The method may be performed by one or more peer appliances in the network. Additionally, steps of the method may be performed in varying orders or concurrently. Furthermore, various steps may be added, removed, or combined in the method and still fall within the scope of the present invention.

In step 510, an appliance retrieves information from the service directory 200. In step 520, the appliance measures performance metric(s) to one or more specified software services using the information from the service directory, such as the IP address or subnet for each software service and test method(s). From the measured metric(s), derived metric(s) are determined for each tested IP address, and the information is stored in the measured metric table 300 at the appliance in step 530. The appliance advertises a selected derived performance metric to the other peer appliances in step 540. As previously disclosed, the appliance may not advertise a derived performance metric if the derived performance metric is outside of a specified threshold. In step 550, the advertised metric table 400 at each peer appliance in the network is updated with the advertised performance metric if an updated advertised performance metric value was advertised. The optimal transit appliance for each software service is determined from the advertised metric table, as discussed above. The advertised metric may also have a time period for which it is valid, upon expiry of which it is calculated, selected, or advertised again.

Each step of the method may be performed at different times (asynchronously), even though it is depicted as a sequence in FIG. 5. For example, if the measured metric doesn't change for a particular appliance, then it may not be advertised in step 540. Additionally, each appliance in the network can perform each step of this method at varying times. Steps of the method may be performed on a periodic fixed schedule, at the direction of a network administration, or upon any other triggering event.

FIG. 6 shows an exemplary system environment suitable for implementing methods for optimization of data across one or more interconnected networks. Three software services are depicted in the exemplary embodiment of FIG. 6, software service 110A, 110B, and 110N. However, there can be any number of software services in communication with the various appliances of the overlay network. While two appliances are depicted in the exemplary system of FIG. 6, there can be any number of appliances in communication over one or more interconnected networks.

In the exemplary system of FIG. 6, an end user accesses a software service, such as software service 110N, through computer 102 by sending data packets to software service 110N via appliance 104 through network 606. As discussed above with respect to FIGS. 1A and 1B, computer 102 may be a desktop computer, laptop computer, handheld computing device, server, or any other type of computing device. While a single computer is depicted here, computer 102 may also be a cluster of computing devices. Also, network 606 may be any type of network, as discussed above with respect to network 108 of FIG. 1B.

Appliance 104 may extract the IP address for software service 110N from the destination IP address in the data packets it receives from the computer 102. Appliance 104 may then query its advertised metric table 400 for the peer appliance in the overlay network via which to direct the request for the software service based on the extracted IP address. The selected peer appliance may constitute the optimal transit appliance for the extracted IP address. If the advertised metric table 400 contains a transit appliance for the extracted IP address of the software service, appliance 104 directs the request via the transit appliance noted for the IP address. A row of the advertised metric table 400 is said to contain an IP address, if that IP address belongs inside the subnet that the row corresponds to. Furthermore, a user request for a software service may be directed through any number of network appliances, routers, switches, or other network devices before the request is routed to the software service, depending on the network path.

In the exemplary embodiment depicted in FIG. 6, appliance 104 determines that the optimal transit appliance for the extracted IP address associated with desired software service 110N is through appliance 106 located at service 118A. Service 118A may contain computing devices that enable software service 110N, or may be unrelated to software service 110N. By placing appliance 106 at IaaS service 118A, appliance 106 may be located close to software service 110N. As such, appliance 106 may have good network performance characteristics for data transfer to and from software service 110N and is likely to be a good transit appliance for software service 110N.

While services 118A, 118B, and 118N are depicted in FIG. 6 as exemplary cloud services within an IaaS location, they may be located outside of the cloud. Furthermore, appliance 106 may be located anywhere in the world, and may not necessarily be in an IaaS center. FIG. 6 depicts an exemplary embodiment where appliance 106 is the selected transit appliance for software service 110N and is located at an IaaS service.

In various embodiments, appliance 106 also performs network address translation (NAT) on the data before forwarding the software service request to software service 110N, such that the request for software service 110N appears to originate from appliance 106. This way, the reply from software service 110N is also routed back through the transit appliance 106.

In some cases, appliance 104 may determine that there is no optimal transit appliance for software service 110N, or the transit appliance is appliance 104. If there is no optimal transit appliance for software service 110N, the user request for software service 110N may be directed from appliance 104 to software service 110N over the network 620, without using the overlay of optimizing peer appliances. Network 620 can be any type of network, including a Wide Area Network (WAN), the Internet, and so forth. In various embodiments, default routing behavior is also stored in one or more routing tables. The routing tables can be stored in each appliance of the network, and/or in a central location accessible to all appliances.

Software service 110N may process the data packets received from appliance 106 and direct the reply to the appliance from which the request was forwarded, in this case appliance 106 located in Service 118A. Appliance 106 then performs network address translation on the data, to direct it to the appliance originating the request, appliance 104. From appliance 104 the reply is sent back to computer 102. In various embodiments, there may be any number of intermediate appliances between appliance 104 and software service 110N. Each intermediate appliance may perform network address translation to ensure that the reply is routed back through the network via the same path.

FIG. 7A is an exemplary flow diagram 700A for the transmission of data packets for a software service through one or more interconnected networks via a first appliance. The method may be performed by one or more peer appliances in the network. Additionally, steps of the method may be performed in varying orders or concurrently. Furthermore, various steps may be added, removed, or combined in the method and still fall within the scope of the present invention.

At step 710, a first appliance (such as appliance 104) receives data packets sent by a user destined for a software service from computer 102. In step 720, the first appliance extracts the destination IP address for the software service from the received data packets. At step 730, the first appliance determines if the extracted destination IP address is in one of the subnets in the advertised metric table 400. If not, the first appliance transmits the data packets destined for the software service to the destination IP address for the software service via default routing behavior in step 740. If the destination IP address is in the advertised metric table 400, the first appliance queries the advertised metric table 400 for the selected transit appliance for the destination IP address, in step 750. While IP addresses are used in this example, the invention can also be applied to other network addressing types.

At step 760A, the first appliance may optionally optimize the data packets destined for the software service. Data optimization techniques may comprise compression/decompression, deduplication, TCP acceleration, performance enhancing proxy, packet reconstruction, error correction, or any other technique for optimizing data transfer between network appliances or devices. For simplification purposes, the term ‘optimization encoding’ is used in the figures. However, a person of ordinary skill in the art would understand that any optimization technique may be applied. Optimization encoding and decoding are symmetric transformations of data, such as compression/decompression, deduplication, etc. For example, data packets that are compressed at a first appliance need to be decompressed at a second appliance. At step 760B, the first appliance transmits the data packets for the software service to the optimal transit appliance with the selected performance metric. Optimization may be performed on a packet by packet basis, such that there is an encoded packet for each original packet, or optimization may be performed on parts of packets or across multiple packets, such that there is not a 1:1 correspondence between the original packets and the encoded packets.

FIG. 7B is an exemplary flow diagram 700B for the transfer of data packets for a software service through a transit appliance, also referred to as a second appliance. At step 770A, a second appliance (such as transit appliance 106 from FIG. 6) receives plain or encoded data packets representing data sent by a user destined for a software service from computer 102. At step 770B, the second appliance optionally applies optimization decoding of the data packets. If data packets from the first appliance were optimized in step 760A in any way, such as encoded, then the packets may be decoded at step 770B.

In step 772, the second appliance performs network address translation to change the source network address in the data packets to its own local network address. At step 774, the second appliance sends the modified data packets to the destination IP address of the requested software service. Response packets are received from the software service at step 776. The second appliance then maps the destination address from the response packets to the original user's IP address (such as the IP address of computer 102), at step 778. The data packets from the software service are then optionally encoded at step 780A by the second appliance. This may be a similar step to the optimization technique applied at the first appliance in step 760A, or a different optimization technique may be applied to the reply data packets. The data packets are transmitted back to the first appliance (or ingress appliance) at step 780B. The first appliance transmits the response data packets from the software service to computer 102.

FIG. 8 illustrates an exemplary screenshot of a graphical user interface (GUI) 800 for a user to select optimization of data traffic to and from particular software services. The GUI 800 may be shown on a display of a user device (not shown) such as a personal computer (PC), a tablet computer, a mobile device, or any other suitable device. In an example, the GUI 800 is shown on the display of the user device via a browser or some other software application.

In various embodiments, the GUI 800 has a listing in column 810 of software services that are available for optimization. The service listing in column 810 may be updated on a periodic fixed schedule, upon the direction of a network administrator, or upon a triggering event. Column 820 of the GUI is an optional column that can show one or more IP addresses or subnet associated with each service. For each service available for optimization, the GUI 800 can optionally also provide the selected transit appliance from the overlay network to the service, in column 830. In column 840, a network administrator or end user can select which service it would like to determine the optimal transit appliance for. In exemplary embodiments, an end user may choose to enable optimization only for services that are actually used, or for services that are used frequently. Even though only checkboxes are shown in the optimization table, other selectable items can be provided, such as radio buttons or the like.

FIG. 9 shows an exemplary global network of appliances in the overlay network. While there are only five appliances depicted in the figure, there can be any number of appliances connected to the overlay network, and they can be located in any geographic location around the world. A request for software service 110A may originate from computer 102 in any appliance location. Each appliance is in communication with portal 114, and maintains a copy of the service directory 200, a measured metric table 300, and an advertised metric table 400. Each appliance also is in communication with the other global appliances, and advertises its performance metric with the peer appliances. Furthermore, each appliance in the network may provide data optimization techniques. The transit appliance for software service 110A may be through any appliance in the global network. In exemplary embodiments, the transit appliance is the appliance geographically located closest to software service 110A, but does not have to be.

Thus, methods and systems for determining a transit appliance for data traffic to and from a software service are disclosed. Although embodiments have been described with reference to specific example embodiments, it will be evident that various modifications and changes can be made to these example embodiments without departing from the broader spirit and scope of the present application. Accordingly, the specification and drawings are to be regarded in an illustrative rather than a restrictive sense.

Claims

1. A computer-implemented method for optimization of network traffic to a software service implemented as a software as a service (SaaS) within an overlay network, the method comprising: receiving network traffic destined for the software service as determined based on a network address contained in the network traffic;determining at least a first and second performance metrics associated with a corresponding network appliance of first and second network appliances and the software service, wherein each performance metric is measured with at least one network address usable to measure the performance metric;selecting the first network appliance over the second network appliance as a destination for the network traffic based on the first and second performance metrics, andreceiving a service directory, the service directory comprising a list of available software services, a corresponding at least one network address for each software service of the available software services, and at least one test method for each combination of the software service and the corresponding at least one network address, wherein the first and second network appliances receive the service directory from a portal and maintain copies of the service directory and update the service directory based on a change to the service directory at the portal.

2. The method of claim 1, wherein the change includes at least one of an addition of a new software service, deletion of a software service, a change in a network address associated with a software service, or a change in a test method associated with a combination of a software service and a network address.

3. The method of claim 1, wherein the first network appliance uses the at least one test method to determine the first performance metric associated with the first network appliance for the software service.

4. The method of claim 3, wherein the first network appliance determines a derived performance metric that is calculated or selected based on the first performance metric.

5. The method of claim 4, wherein the determining the at least the first and second performance metrics associated with a corresponding network appliance of the first and second network appliances comprises: receiving the derived performance metric from the first network appliance.

6. The method of claim 4, further comprising: maintaining an advertised metric table comprising the first and second network appliances and advertised metrics received from corresponding network appliances of the first and second network appliances; andquerying the advertised metric table to determine the first and second performance metrics.

7. The method of claim 6, wherein the advertised metric table organizes the first and second network appliances according to one or more associated subnets.

8. A system for optimization of network traffic to a software service implemented as a software as a service (SaaS) within an overlay network, the system comprising: at least one processor; anda memory storing instructions that, when executed by the at least one processor, cause the system to perform a method comprising:receiving network traffic destined for the software service as determined based on a network address contained in the network traffic;determining at least a first and second performance metrics associated with a corresponding network appliance of first and second network appliances and the software service, wherein each performance metric is measured with at least one network address usable to measure the performance metric;selecting the first network appliance over the second network appliance as a destination for the network traffic based on the first and second performance metrics, andreceiving a service directory, the service directory comprising a list of available software services, a corresponding at least one network address for each software service of the available software services, and at least one test method for each combination of the software service and the corresponding at least one network address, wherein the first and second network appliances receive the service directory from a portal and maintain copies of the service directory and update the service directory based on a change to the service directory at the portal.

9. The system of claim 8, wherein the change includes at least one of an addition of a new software service, deletion of a software service, a change in a network address associated with a software service, or a change in a test method associated with a combination of a software service and a network address.

10. The system of claim 8, wherein the first network appliance uses the at least one test method to determine the first performance metric associated with the first network appliance for the software service.

11. The system of claim 10, wherein the first network appliance determines a derived performance metric that is calculated or selected based on the first performance metric.

12. The system of claim 11, wherein the memory storing instructions that, when executed by the at the least one processor, cause the system to perform the method comprising determining the at least the first and second performance metrics associated with a corresponding network appliance of the first and second network appliances, further cause the system to perform the method comprising: receiving the derived performance metric from the first network appliance.

13. The system of claim 11, the memory storing instructions that, when executed by the at least one processor, cause the system to perform the method comprising: maintaining an advertised metric table comprising the first and second network appliances and advertised metrics received from corresponding network appliances of the first and second network appliances; andquerying the advertised metric table to determine the first and second performance metrics.

14. The system of claim 13 wherein the advertised metric table organizes the first and second network appliances according to one or more associated subnets.

15. A non-transitory computer-readable storage medium including instructions that, when executed by at least one processor of a computing system, cause the computing system to perform a method for optimization of network traffic to a software service implemented as a software as a service (SaaS) within an overlay network, the method comprising: receiving network traffic destined for the software service as determined based on a network address contained in the network traffic;determining at least a first and second performance metrics associated with a corresponding network appliance of first and second network appliances and the software service, wherein each performance metric is measured with at least one network address usable to measure the performance metric;selecting the first network appliance over the second network appliance as a destination for the network traffic based on the first and second performance metrics, andreceiving a service directory, the service directory comprising a list of available software services, a corresponding at least one network address for each software service of the available software services, and at least one test method for each combination of the software service and the corresponding at least one network address, wherein the first and second network appliances receive the service directory from a portal and maintain copies of the service directory and update the service directory based on a change to the service directory at the portal.

16. The non-transitory computer-readable storage medium of claim 6, wherein the change includes at least one of an addition of a new software service, deletion of a software service, a change in a network address associated with a software service, or a change in a test method associated with a combination of a software service and a network address.

17. The non-transitory computer-readable storage medium of claim 15, wherein the first network appliance uses the at least one test method to determine the first performance metric associated with the first network appliance for the software service.

18. The non-transitory computer-readable storage medium of claim 17, wherein the first network appliance determines a derived performance metric that is calculated or selected based on the first performance metric.

19. The non-transitory computer-readable storage medium of claim 18 storing instructions that, when executed by the at least one processor of the computing system, cause the computing system to perform the method comprising, determining the at least the first and second performance metrics associated with a corresponding network appliance of the first and second network appliances, further cause the computing system to perform the method comprising: receiving the derived performance metric from the first network appliance.

20. The non-transitory computer-readable storage medium of claim 18 storing instructions that, when executed by the at least one processor of the computer system, cause the computing system to perform the method comprising: maintaining an advertised metric table comprising the first and second network appliances and advertised metrics received from corresponding network appliances of the first and second network appliances; andquerying the advertised metric table to determine the first and second performance metrics.