DNS-based GSLB-aware SD-WAN for low latency SaaS applications

Abstract

Some embodiments of the invention provide a method of sending data messages from an edge router at a first location of an enterprise network to a SaaS (software as a service) application server provided by a third-party at a second location. The method receives, from a DNS (domain name system) first server, a resolution for a particular destination network address for the SaaS application server at the second location. From a second server, the method obtains an identifier for a first cloud gateway from multiple cloud gateways at multiple locations through which the particular destination address for the SaaS application server can be reached, the first cloud gateway farther from the first location than a second cloud gateway in the multiple cloud gateways but closer to the second location than the second cloud gateway. The method uses an optimized SD-WAN connection to the first cloud gateway to forward data messages for the first cloud gateway to the SaaS application at the second location.

Description

BACKGROUND

Today, SD-WAN solutions can provide different networking services according to defined business policies, such as the VMware SD-WAN by VeloCloud®. These networking services can include direct internet, internet via a cloud gateway (e.g., VeloCloud® gateway), and internet via private link (e.g., MPLS) backhaul. Typically, the business policy and default policy direct traffic to the internet via a cloud gateway. However, while this approach may direct traffic to a specific cloud gateway based on proximity to the source location, this specific cloud gateway may not be the best gateway for certain publicly hosted applications, such as applications where the specific cloud gateway is unable to meet low latency requirements of the application.

BRIEF SUMMARY

Some embodiments of the invention provide a method of routing data message traffic between an edge router at a branch first location for an enterprise network and a SaaS (software as a service) application server provided by a third-party at a second location. The method is performed, in some embodiments, by the edge router at the branch first location. The method queries a (global server load balancing) GSLB-aware DNS (domain name system) server for a cloud gateway from multiple cloud gateways at multiple locations through which a particular destination network address for the SaaS application server can be reached. From the GSLB-aware DNS server, the method receives an identifier for a first cloud gateway that is farther from the branch first location than a second cloud gateway, but closer to the second location than the second cloud gateway. The method then uses an optimized SD-WAN connection to the first cloud gateway to forward data messages for the first cloud gateway to the SaaS application at the second location.

In some embodiments, each link between the edge router and a cloud gateway is associated with a respective latency score. For instance, a first link between the edge router and the first cloud gateway is associated with a first latency score, in some embodiments, and a second link between the edge router and the second cloud gateway. In some embodiments, after receiving the identifier for the first cloud gateway, the edge router sends requests to the first and second cloud gateways for latency scores associated with the respective links between the first and second cloud gateways and the SaaS application server. In response to the requests, the edge router of some embodiments receives a third latency score from the first cloud gateway associated with a third link between the first cloud gateway and the SaaS application server, and receives a fourth latency score from the second cloud gateway associated with a fourth link between the second cloud gateway and the SaaS application server.

The first latency score associated with the first link between the edge router and the first cloud gateway and the fourth latency score associated with the fourth link between the second cloud gateway and the SaaS application server are high latency scores, in some embodiments. Conversely, in some embodiments, the second latency score associated with the second link between the edge router and the second cloud gateway, and the third latency score associated with the third link between the first cloud gateway and the SaaS application server are low latency scores. In some embodiments, the SaaS application server is associated with a low latency requirement that specifies a maximum latency threshold for the last mile connectivity to the SaaS application (i.e., the unmanaged, unoptimized links from the cloud gateways to the SaaS application server). In some such embodiments, the first cloud gateway is identified as the optimal cloud gateway for reaching the SaaS application server based on a determination that the third latency score meets the low latency requirement (i.e., does not exceed the maximum latency threshold), while the fourth latency score does not meet the low latency requirement (i.e., does exceed the maximum latency threshold).

In some embodiments, the first and second cloud gateways identify the latency scores provided to the edge router by sending probe packets to the SaaS application server. The first and second cloud gateways derive latency measurements from the probe packets to generate the latency scores for the links between the first and second cloud gateways and the SaaS application server, according to some embodiments. The probe packets, in some embodiments, include layer 4 (L4) and layer 7 (L7) probe packets.

In addition to the first cloud gateway being farther from the first location than the second cloud gateway, the first cloud gateway of some embodiments is also associated with a longer (round-trip time) RTT than the second cloud gateway (i.e., due to the increased distance to each the first cloud gateway). In some embodiments, the first cloud gateway is farther from the edge router than the second cloud gateway in terms of physical distance, while in other embodiments, the first cloud gateway is farther from the edge router than the second cloud gateway in terms of signal distance. In still other embodiments, the first cloud gateway is farther from the edge router than the second cloud gateway in terms of both physical distance and signal distance. The edge router of some embodiments determines that the first cloud gateway is farther than the second cloud gateway based on proximity information provided by a server (e.g., the DNS server or a management and control server), while in other embodiments, the edge router makes this determination by performing a proximity operation to determine the distance to each cloud gateway.

In some embodiments, the DNS server is a DNS first server that provides a particular destination network address for the SaaS application server to the edge router, while a second server provides the identifier for the first cloud gateway to the edge router. The second server, in some embodiments, is a management and control server for the enterprise network that identifies the most optimal cloud gateways for each edge router in the enterprise network to reach the SaaS application server, and provides identifiers for the most optimal cloud gateways to each edge router.

The second server, in some embodiments, is also responsible for directing each cloud gateway to compute latency scores (i.e., rather than the edge routers each directing the cloud gateways to compute latency scores), and periodically collects the latency scores to dynamically identify the most optimal cloud gateways for each edge router. In some embodiments, the second server causes the edge router to automatically switch from the first cloud gateway to another cloud gateway upon identifying the other gateway router as the most optimal cloud gateway based on updated latency scores for the first cloud gateway and the other cloud gateway. For instance, in some embodiments, the second cloud gateway is a default gateway assigned to the edge router for forwarding data messages to the SaaS application server until the edge router receives the identifier for the first cloud gateway when the second server determines that the first cloud gateway is the most optimal gateway for use by the edge router to reach the SaaS application server. The second server of some embodiments also provides the location information and latency scores associated with the first cloud gateway to the edge router when providing the identifier for the first cloud gateway.

The preceding Summary is intended to serve as a brief introduction to some embodiments of the invention. It is not meant to be an introduction or overview of all inventive subject matter disclosed in this document. The Detailed Description that follows and the Drawings that are referred to in the Detailed Description will further describe the embodiments described in the Summary as well as other embodiments. Accordingly, to understand all the embodiments described by this document, a full review of the Summary, the Detailed Description, the Drawings, and the Claims is needed. Moreover, the claimed subject matters are not to be limited by the illustrative details in the Summary, the Detailed Description, and the Drawings.

BRIEF DESCRIPTION OF FIGURES

The novel features of the invention are set forth in the appended claims. However, for purposes of explanation, several embodiments of the invention are set forth in the following figures.

FIG. 1 conceptually illustrates a block diagram of different routes between a branch office of an enterprise network for an entity and a geographically distant datacenter hosting a SaaS application server, in some embodiments.

FIG. 2 conceptually illustrates a process of some embodiments for identifying a cloud gateway for forwarding data messages to a datacenter hosting a SaaS application server.

FIG. 3, for instance, conceptually illustrates the block diagram of FIG. 1 with latency scores of some embodiments associated with the connections between the LA branch office, cloud gateways, and Japan datacenter.

FIG. 4 conceptually illustrates a block diagram of a workflow of some embodiments for a branch office to identify a cloud gateway from a set of cloud gateways based on latency scores associated with using each cloud gateway in the set.

FIG. 5 conceptually illustrates a process for identifying a cloud gateway for forwarding data messages to a SaaS application server, in some embodiments.

FIG. 6 conceptually illustrates a block diagram of some embodiments in which a branch office obtains cloud gateway information from a controller for use in forwarding data messages to a SaaS application server.

FIG. 7 conceptually illustrates a process of some embodiments for obtaining a cloud gateway from a controller for use in forwarding data messages to a SaaS application server.

FIG. 8 illustrates an example of a standard SD-WAN environment in which some embodiments of the invention are implemented.

FIG. 9 conceptually illustrates a computer system with which some embodiments of the invention are implemented.

DETAILED DESCRIPTION

In the following detailed description of the invention, numerous details, examples, and embodiments of the invention are set forth and described. However, it will be clear and apparent to one skilled in the art that the invention is not limited to the embodiments set forth and that the invention may be practiced without some of the specific details and examples discussed.

Some embodiments of the invention provide a method of routing data message traffic between an edge router at a branch first location for an enterprise network and a SaaS (software as a service) application server provided by a third-party at a second location. The method is performed, in some embodiments, by the edge router at the branch first location. The method queries a (global server load balancing) GSLB-aware DNS (domain name system) server for a cloud gateway from multiple cloud gateways at multiple locations through which a particular destination network address for the SaaS application server can be reached. From the GSLB-aware DNS server, the method receives an identifier for a first cloud gateway that is farther from the first location than a second cloud gateway, but closer to the second location than the second cloud gateway. The method then uses an optimized SD-WAN connection to the first cloud gateway to forward data messages for the first cloud gateway to the SaaS application at the second location.

In some embodiments, each link between the edge router and a cloud gateway is associated with a respective latency score. For instance, a first link between the edge router and the first cloud gateway is associated with a first latency score, in some embodiments, and a second link between the edge router and the second cloud gateway. In some embodiments, after receiving the identifier for the first cloud gateway, the edge router sends requests to the first and second cloud gateways for latency scores associated with the respective links between the first and second cloud gateways and the SaaS application server. In response to the requests, the edge router of some embodiments receives a third latency score from the first cloud gateway associated with a third link between the first cloud gateway and the SaaS application server, and receives a fourth latency score from the second cloud gateway associated with a fourth link between the second cloud gateway and the SaaS application server.

The first latency score associated with the first link between the edge router and the first cloud gateway and the fourth latency score associated with the fourth link between the second cloud gateway and the SaaS application server are high latency scores, in some embodiments. Conversely, in some embodiments, the second latency score associated with the second link between the edge router and the second cloud gateway and the third latency score associated with the third link between the first cloud gateway and the SaaS application server are low latency scores. In some embodiments, the SaaS application server is associated with a low latency requirement that specifies a maximum latency threshold for the last mile connectivity to the SaaS application (i.e., the unmanaged, unoptimized links from the cloud gateways to the SaaS application server). In some such embodiments, the first cloud gateway is identified as the optimal cloud gateway for reaching the SaaS application server based on a determination that the third latency score meets the low latency requirement (i.e., does not exceed the maximum latency threshold), while the fourth latency score does not meet the low latency requirement (i.e., does exceed the maximum latency threshold).

In some embodiments, the first and second cloud gateways identify the latency scores provided to the edge router by sending probe packets to the SaaS application server. The first and second cloud gateways derive latency measurements from the probe packets to generate the latency scores for the links between the first and second cloud gateways and the SaaS application server, according to some embodiments. The probe packets, in some embodiments, include layer 4 (L4) and layer 7 (L7) probe packets.

In addition to the first cloud gateway being farther from the first location than the second cloud gateway, the first cloud gateway of some embodiments is also associated with a longer (round-trip time) RTT than the second cloud gateway (i.e., due to the increased distance to the first cloud gateway). In some embodiments, the first cloud gateway is farther from the edge router than the second cloud gateway in terms of physical distance, while in other embodiments, the first cloud gateway is farther from the edge router than the second cloud gateway in terms of signal distance. In still other embodiments, the first cloud gateway is farther from the edge router than the second cloud gateway in terms of both physical distance and signal distance. The edge router of some embodiments determines that the first cloud gateway is farther than the second cloud gateway based on proximity information provided by a server (e.g., the DNS server or a management and control server), while in other embodiments, the edge router makes this determination by performing a proximity operation to determine the distance to each cloud gateway.

In some embodiments, the DNS server is a DNS first server that provides a particular destination network address for the SaaS application server to the edge router, while a second server provides the identifier for the first cloud gateway to the edge router. The second server, in some embodiments, is a management and control server for the enterprise network that identifies the most optimal cloud gateways for each edge router in the enterprise network to reach the SaaS application server, and provides identifiers for the most optimal cloud gateways to each edge router.

The second server, in some embodiments, is also responsible for directing each cloud gateway to compute latency scores (i.e., rather than the edge routers each directing the cloud gateways to compute latency scores), and periodically collects the latency scores to dynamically identify the most optimal cloud gateways for each edge router. In some embodiments, the second server causes the edge router to automatically switch from the first cloud gateway to another cloud gateway upon identifying the other gateway router as the most optimal cloud gateway based on updated latency scores for the first cloud gateway and the other cloud gateway. For instance, in some embodiments, the second cloud gateway is a default gateway assigned to the edge router for forwarding data messages to the SaaS application server until the edge router receives the identifier for the first cloud gateway when the second server determines that the first cloud gateway is the most optimal gateway for use by the edge router to reach the SaaS application server. The second server of some embodiments also provides the location information and latency scores associated with the first cloud gateway to the edge router when providing the identifier for the first cloud gateway.

FIG. 1 conceptually illustrates a block diagram of different routes between a branch office of an enterprise network for an entity and a geographically distant datacenter hosting a SaaS application server, in some embodiments. As shown, the diagram 100 includes a Los Angeles (“LA”) branch office 110 and a Japan datacenter 120 at which a SaaS application server is located. The LA branch office 110 is connected through an SD-WAN (e.g., to cloud gateways 130 and 135 through which data messages to the Japan datacenter 120 can be forwarded. While only one branch office 110 is shown, the SD-WAN (also referred to herein as a virtual network) may connect multiple branch sites to each other, as well as to, e.g., a controller (not shown).

The connections between the LA branch office 110 and each of the cloud gateways 130-135 are between an edge router (e.g., VeloCloud® edge (VCE)) at the branch office and the cloud gateways (e.g., VeloCloud® gateways (VCGs)), according to some embodiments. The edge routers of some embodiments are edge machines (e.g., virtual machines (VMs), containers, programs executing on computers, etc.) and/or standalone appliances that operate at multi-computer locations of the particular entity (e.g., at an office or datacenter of the entity) to connect the computers at their respective locations to other nodes (e.g., gateways, hubs, etc.) in the virtual network. In some embodiments, the nodes are clusters of nodes at each of the branch sites. In other embodiments, the edge nodes are deployed to each of the branch sites as high-availability pairs such that one edge node in the pair is the active node and the other edge node in the pair is the standby node that can take over as the active edge node in case of failover.

An example of an entity for which such a virtual network can be established includes a business entity (e.g., a corporation), a non-profit entity (e.g., a hospital, a research organization, etc.), and an education entity (e.g., a university, a college, etc.), or any other type of entity. Examples of public cloud providers include Amazon Web Services® (AWS), Google Cloud Platform™ (GCP), Microsoft Azure®, etc., while examples of entities include a company (e.g., corporation, partnership, etc.), an organization (e.g., a school, a non-profit, a government entity, etc.), etc. In other embodiments, hub forwarding elements can also be deployed in private cloud datacenters of a virtual WAN provider that hosts these hubs to establish SD-WANs for different entities.

In this example, one cloud gateway 130 is located in California and the other cloud gateway 135 is located in Japan. The LA branch office 110 has two route options using the two cloud gateways for reaching the Japan datacenter 120. The first route includes connection 1A from the LA branch office 110 to the California cloud gateway 130 and connection 1B from the California cloud gateway 130 to the Japan datacenter 120. The second route includes connection 2A from the LA branch office 110 to the Japan cloud gateway 135 and connection 2B from the Japan cloud gateway 135 to the Japan datacenter 120. The respective connections 1A and 2A from the LA branch office 110 to the California and Japan cloud gateways 130 and 135 are managed by the SD-WAN and therefore can be optimized. In other words, when both forwarding elements (i.e., a forwarding element at the LA branch office 110 and either of the California and Japan cloud gateways 130 and 135) for one hop (i.e., source and destination forwarding elements) are managed by the SD-WAN, controllers for the SD-WAN can configure the two managed forwarding elements to perform their forwarding operations in a way that maximizes QoS (quality of service), or other metrics such as throughput, according to some embodiments.

In some embodiments, for instance, multiple links of any type (e.g., DSL (digital subscriber line), cable modem, Ethernet, fiber, LTE/3G/4G/5G, MPLS (multiprotocol label switching), radio, satellite, Wi-Fi, etc.), location, or provider can be leveraged to optimize connections between branch offices and cloud gateways. Also, in some embodiments, managed links can be configured to provide a desirable attribute associated with certain metrics such as throughput, QoS, and packet drops. Additional methods of link optimization, in some embodiments, can include utilizing link scores (e.g., assigned based on QoS metrics collected from devices connected by the SD-WAN), alternate path selections, packet replication, forward error correction, etc. For example, some embodiments utilize dynamic multipath optimization (DMPO) to optimize links in the SD-WAN. DMPO, according to some embodiments, is a cloud-delivered architecture for on-premises and cloud applications that enables the use of multiple WAN transports simultaneously, thereby maximizing bandwidth and ensuring application performance.

Returning to the diagram 100, while the distance from the LA branch office 110 to the Japan cloud gateway 135 is longer than the distance from the LA branch office 110 to the California cloud gateway 130, the last mile connectivity from the Japan cloud gateway 135 to the Japan datacenter 120 is shorter than the last mile connectivity from the California cloud gateway 130 to the Japan datacenter 120. Unlike the connections 1A and 2A between the LA branch office 110 and the California and Japan cloud gateway 130 and 135, the last mile connectivity links 1B and 2B are not managed by the SD-WAN and do not benefit from the same optimizations as the links that are managed by the SD-WAN. As a result, the second route through the Japan cloud gateway 135 is the optimal route due to the distance covered by the unmanaged portion of the path, according to some embodiments. That is, while the first route through the California cloud gateway 130 may be associated with better overall metrics (e.g., shorter round-trip time), the combination of the potential for optimizing the link covering the greater distance (i.e., link 2A) and the shorter length of link 2B compared to link 1B that are associated with using the Japan cloud gateway 135 make the second route through the Japan cloud gateway 135 the optimal route (i.e., compared to the route through California cloud gateway 130).

FIG. 2 conceptually illustrates a process 200 of some embodiments for identifying a cloud gateway for forwarding data messages to a datacenter hosting a SaaS application server. The process 200 is performed in some embodiments by an edge router at a branch office of an enterprise network (e.g., an edge router at the LA branch office 110 in the diagram 100). The process 200 starts when the edge router identifies (at 210) a destination datacenter in which a particular SaaS application is located. In some embodiments, the edge router identifies the destination datacenter by requesting a network address (e.g. IP address) associated with the SaaS application from a DNS server, and receiving a network address and location information in response to the request.

From a set of cloud gateways through which the destination datacenter hosting the SaaS application can be reached, the process 200 identifies (at 220) the most optimal cloud gateway for forwarding data messages to the destination datacenter. In some embodiments, the edge router identifies the most optimal cloud gateway based on receiving an identifier for a particular cloud gateway from the DNS server from which the network address and location of the destination datacenter were received. In other embodiments, the edge router identifies the most optimal cloud gateway based on receiving an identifier for a particular cloud gateway from a management and control server. In still other embodiments, the edge router compares location and latency information associated with the set of cloud gateways to identify the most optimal cloud gateway for itself. For instance, the edge router of some embodiments determines whether a default (e.g., primary) cloud gateway assigned to the edge router and the gateway nearest to the DNS server match, and when these cloud gateways do not match, the edge router identifies the cloud gateway with the lowest latency to the SaaS application server.

In some embodiments, the edge router obtains latency data from the cloud gateways by directing the cloud gateways to perform probes to the SaaS application server and to compute latency scores based on those probes for providing to the edge router. The edge router of some embodiments also already has latency scores for connections from the edge router to the cloud gateways based on DMPO. As mentioned above, DMPO, according to some embodiments, is a cloud-delivered architecture for on-premises and cloud applications that enables the use of multiple WAN transports simultaneously, thereby maximizing bandwidth and ensuring application performance.

Returning to the process 200, the process 200 then uses (at 230) the identified cloud gateway to forward data messages to the SaaS application located at the destination datacenter. In the diagram 100, for instance, an edge router at the LA branch office 110 would use an optimized SD-WAN connection to either the California cloud gateway 130 or the Japan cloud gateway 135 based on which cloud gateway has been identified as the most optimal cloud gateway to reach the SaaS application server at the Japan datacenter 120. Following 230, the process 200 ends.

In some embodiments, the cloud gateway used by the edge router to forward data messages to a SaaS server is selected based on latency scores associated with the links between the edge router, cloud gateways, and destination datacenter. FIG. 3, for instance, conceptually illustrates the block diagram 100 of FIG. 1 with latency scores of some embodiments associated with the connections between the LA branch office 110, cloud gateways 130-135, and Japan datacenter 120.

As shown, the link 340 between the LA branch office 110 and the California cloud gateway 130 and the link 355 between the Japan cloud gateway 135 are indicated as being low latency links. Conversely, the link 345 between the California cloud gateway 130 and the Japan datacenter 120 and the link 350 between the LA branch office 110 and the Japan cloud gateway 135 are indicated as being high latency links. Thus, the first route from the LA office 110 to the Japan datacenter 120 through the California cloud gateway 130 includes a low latency first connection to the cloud gateway and a high latency second connection for the last mile connectivity from the cloud gateway 130 to the datacenter 120, while the second route from the LA office 110 to the Japan datacenter 120 through the Japan cloud gateway 135 includes a high latency first connection to the cloud gateway 135 and a low latency second connection for the last mile connectivity from the cloud gateway 135 to the Japan datacenter 120.

In some embodiments, the SaaS application at the Japan datacenter 120 has low latency requirements, and the Japan cloud gateway 135 is the more optimal cloud gateway compared to the California cloud gateway 130 based on the connection between the Japan cloud gateway 135 and the Japan datacenter 120 having low latency, while the connection between the California cloud gateway 130 and Japan datacenter 120 has high latency. In some embodiments, the connection between the Japan cloud gateway 135 and the Japan datacenter 120 has lower latency than the connection between the California cloud gateway 130 and the Japan datacenter 120 based on the shorter distance between the Japan cloud gateway 135 and Japan datacenter 120.

FIG. 4 conceptually illustrates a block diagram of a workflow of some embodiments for a branch office to identify a cloud gateway from a set of cloud gateways based on latency scores associated with using each cloud gateway in the set. As illustrated, the diagram 400 includes an LA branch office 410, a Japan datacenter 420 that hosts a SaaS application server, a cloud gateway 430 located in California, a cloud gateway 435 located in Japan, a DNS server 440, and location-based context 445.

The diagram 400 will be further described below by reference to FIG. 5, which conceptually illustrates a process 500 for identifying a cloud gateway for forwarding data messages to a SaaS application server, in some embodiments. The process 500 is performed in some embodiments by an edge router for a branch office of an enterprise network. The process 500 starts when the edge router queries (at 510) a DNS server for a cloud gateway for forwarding data messages to a SaaS application server. Like the diagram 400, the SaaS application server of some embodiments is located in a datacenter that is geographically distant from the branch office trying to reach the SaaS application server. The edge router at the branch office 410 sends a request for a cloud gateway (at the encircled 1) to the DNS server 440.

In some embodiments, prior to requesting the cloud gateway, the edge router first requests a network address and location information for the SaaS application server. The DNS server 440 of some embodiments obtains location information for the SaaS application server at the resolved network address from the location-based context 445, and provides both the network address and location information to the edge router at the branch office 410 in response to the request.

The process 500 receives (at 520) an identifier for a cloud gateway from the DNS server. For instance, the branch office 410 receives (at the encircled 2) an identifier for the Japan cloud gateway 435. In some embodiments, prior to requesting and receiving the cloud identifier, the edge router is assigned a default cloud gateway for use in forwarding data messages to third-party SaaS application servers, and thus already has an identifier for at least one other cloud gateway. In other embodiments, the DNS server responds to the edge router's request by providing two or more identifiers to the edge router for the edge router to select from.

The process 500 sends (at 530) a request for a latency score associated with a connection between the cloud gateway and SaaS application server to the cloud gateway. The SaaS application server of some embodiments is associated with a low latency requirement, and thus the edge router of some such embodiments sends a request to the cloud gateway for a latency score to determine whether the latency score meets the low latency requirement. In the diagram 400, for instance, the LA branch office 410 (i.e., an edge router at the branch office) sends requests for latency scores (at the encircled 3) to the California and Japan cloud gateways 430-435.

In some embodiments, the cloud gateways compute latency scores by sending probe packets to the SaaS application server and collecting latency measurements using the probe packets. For example, the California and Japan cloud gateways 430-435 are each shown (at the encircled 4) performing probes for latency measurements between the cloud gateways 430-435 and the Japan datacenter 420 that hosts the SaaS application server. The probe packets, in some embodiments, include L4 and L7 probe packets.

The process 500 then receives (at 540) a latency score from the cloud gateway, such as the latency scores received (at the encircled 5) at the LA branch office 410 in the diagram 400, and determines (at 550) whether the latency score meets the low latency requirement associated with the SaaS application server. In some embodiments, the distance between a cloud gateway and the destination SaaS application server affects latency such that a cloud gateway that is farther from the SaaS application server has a higher latency than a cloud gateway that is closer to the SaaS application server. When the latency score from the cloud gateway does not meet the low latency requirement (e.g., is higher than a specified latency threshold), the process 500 returns to 510 to query the DNS server for a cloud gateway. In some embodiments, the edge router uses or continues to use a default cloud gateway to forward data messages to the SaaS application server while awaiting the next cloud gateway from the DNS server.

When the latency score does meet the low latency requirement (e.g., is lower than a specified latency threshold), the process 500 transitions to use (at 560) an optimized SD-WAN connection to the cloud gateway to forward data messages for the cloud gateway to send to the SaaS application server. Following 560, the process 500 ends.

In some embodiments, rather than obtaining a cloud gateway (i.e., identifier for a cloud gateway) from the DNS server, the edge router instead requests the cloud gateway from a controller for the enterprise network (e.g., a VeloCloud® Orchestrator (VCO)). FIG. 6 conceptually illustrates a block diagram 600 of some embodiments in which a branch office obtains cloud gateway information from a controller for use in forwarding data messages to a SaaS application server.

The controller 650, in some embodiments, manages business policies for the SD-WAN. The controller 650 is hosted by a cloud, in some embodiments, and on-premises in other embodiments. In some embodiments, new business policies may be created (e.g., network preferred DNS) that may trigger different actions, such as configuring or choosing a business policy for a tenant to be implemented by edge routers at the edges of the tenant networks. For example, an edge router of some embodiments may receive a policy specified for a SaaS application, and the edge router then configures a rule to match the business policy. The edge router of some such embodiments may send data messages to each cloud gateway assigned to the edge router to instruct the cloud gateways to perform L4 and/or L7 probes to the SaaS application to collect latency measurements. In other embodiments, the controller 650 directs the cloud gateways to perform these probes and provides the results to the edge router. As mentioned above, the edge router of some embodiments already has latency data from the edge router to each cloud gateway based on DMPO, and also either stores, or obtains, proximity location information for each cloud gateway.

As shown, rather than an edge router at the branch office 610 requesting latency scores from the cloud gateways 630-635, the controller 650 collects latency measurements from the California and Japan cloud gateways 630-635. In some embodiments, the controller 650 collects latency measurements periodically to ensure that each edge router is using the most optimal cloud gateway for forwarding data messages to destinations outside of the enterprise network, such as the third-party SaaS application server at the Japan datacenter 620. For instance, if a particular cloud gateway's updated latency score increases, the controller 650 of some embodiments may cause an edge router assigned to that cloud gateway to automatically switch to a different cloud gateway that is associated with a lower latency score.

In some embodiments, the controller 650 may assign a default cloud gateway to an edge router for reaching third-party sites within a first region, and assign a secondary cloud gateway to the edge router for reaching third-party sites within a second region. For example, the controller 650 of some embodiments may assign the California cloud gateway 630 as a default cloud gateway to an edge router at the LA branch office 610 for use in reaching third-party sites in California and surrounding states, and subsequently assign the Japan cloud gateway 635 as a secondary cloud gateway for use in reaching at least the Japan datacenter 620. The edge router of some such embodiments may automatically switch between the different cloud gateways based on the destination network addresses of data messages being forwarded.

FIG. 6 will be further described below by reference to FIG. 7, which conceptually illustrates a process 700 of some embodiments for obtaining a cloud gateway from a controller for use in forwarding data messages to a SaaS application server. The process 700 is performed in some embodiments by an edge router at a branch office (e.g., LA branch office 610). The process 700 starts when the edge router queries (at 710) a controller for a cloud gateway for forwarding data messages to a SaaS application server.

In some embodiments, for example, after the edge router receives a destination network address and location information for a SaaS application, the edge router may use the received destination network address to query the controller for a cloud gateway to use to reach the SaaS application. In the diagram 600, the LA branch office 610, for example, receives a destination IP for the Japan SaaS application from the DNS server 640 (at the encircled 1), and then receives a cloud gateway identifier and location and latency data for the cloud gateway from the controller 650 (at the encircled 2).

The process 700 receives (at 720) an identifier for a cloud gateway from the controller. In some embodiments, the edge router is assigned a set of cloud gateways by the controller for use in forwarding data messages to entities outside of the enterprise network, and queries the controller for an identifier corresponding to a cloud gateway from the set that is most optimal for reaching the destination IP associated with the SaaS application. As also described above, the controller of some embodiments determines the most optimal cloud gateway for use by an edge router based on latency scores associated with connections between the cloud gateways and the destination.

For example, in the diagram 600, both available routes from the LA branch office 610 to the Japan datacenter 620 include a high latency connection and a low latency connection. However, the low latency connection for the route that uses the Japan cloud gateway 635 is the unmanaged last mile connection (i.e., unoptimized connection) to the Japan datacenter 620, whereas the low latency connection for the route that uses the California cloud gateway 630 is the managed connection (i.e., optimized connection) from the LA branch office 610 to the California cloud gateway 630. While the route through the California cloud gateway 630 may be identified as the best route (e.g., in a routing table of the edge router based on proximity to the LA branch office 610), the Japan cloud gateway 635 may be identified by the controller 650 as the most optimal cloud gateway based on the low latency associated with the unmanaged, unoptimized connection between the Japan cloud gateway 635 and Japan datacenter 620.

The process 700 uses (at 730) an optimized SD-WAN connection to the cloud gateway to forward data messages for the cloud gateway to send to the SaaS application server. The LA branch office 610, for instance, is illustrated (at the encircled 3) as forwarding data messages to the Japan datacenter 620 via the Japan cloud gateway 635. Following 730, the process 700 ends.

FIG. 8 illustrates an example of a standard SD-WAN environment 800 in which some embodiments of the invention are implemented. The SD-WAN 800 is created in some embodiments for a particular entity using SD-WAN forwarding elements (FEs) deployed at branch sites, datacenters, and public clouds. Examples of public clouds are public clouds provided by Amazon Web Services (AWS), Google Cloud Platform (GCP), Microsoft Azure, etc., while examples of entities include a company (e.g., corporation, partnership, etc.), an organization (e.g., a school, a non-profit, a government entity, etc.), etc.

In the SD-WAN 800, the SD-WAN FEs include cloud gateway 805 (e.g., cloud gateway router) and SD-WAN FEs 830, 832, 834, 836. The cloud gateway (CGW) in some embodiments is an FE that is in a private or public datacenter 810. The CGW 805 in some embodiments has secure connection links (e.g., tunnels) with edge FEs (e.g., SD-WAN edge FEs 830, 832, 834, and 836) at the particular entity's multi-machine sites (e.g., SD-WAN edge sites 820, 822, and 824), such as multi-user compute sites (e.g., branch offices or other physical locations having multi user computers and other user-operated devices and serving as source computers and devices for requests to other machines at other sites), datacenters (e.g., locations housing servers), etc. These multi-machine sites are often at different physical locations (e.g., different buildings, different cities, different states, etc.).

Four multi-machine sites 820-826 are illustrated in the SD-WAN 800, with three of them being branch sites 820-824, and one being a datacenter 826. Each branch site is shown to include a respective edge FE 830, 832, and 834 (e.g., edge routers) and respective resources 850, 852, and 854. The resources 850-854, in some embodiments, are machines located at the branch sites 820-824. The datacenter site 826 is shown to include a hub FE 836 (e.g., a hub router), and datacenter resources 856. The datacenter SD-WAN FE 836 is referred to as a hub FE because in some embodiments this FE can be used to connect to other edge FEs of the branch sites 820-824. The hub FE in some embodiments uses or has one or more service engines to perform services (e.g., middlebox services) on packets that it forwards from one branch site to another branch site. The hub FE also provides access to the datacenter resources 856.

Each edge FE (e.g., SD-WAN edge FEs 830-834) exchanges packets with one or more cloud gateways 805 through one or more connection links 815 (e.g., multiple connection links available at the edge FE). In some embodiments, these connection links include secure and unsecure connection links, while in other embodiments they only include secure connection links. As shown by edge FE 834 and gateway 805, multiple secure connection links (e.g., multiple secure tunnels that are established over multiple physical links) can be established between one edge FE and a gateway.

When multiple such links are defined between an edge FE and a gateway, each secure connection link in some embodiments is associated with a different physical network link between the edge FE and an external network. For instance, to access external networks, an edge FE in some embodiments has one or more commercial broadband Internet links (e.g., a cable modem, a fiber optic link) to access the Internet, an MPLS (multiprotocol label switching) link to access external networks through an MPLS provider's network, a wireless cellular link (e.g., a 5G LTE network), etc. In some embodiments, the different physical links between the edge FE 834 and the cloud gateway 805 are the same type of links (e.g., are different MPLS links).

In some embodiments, one edge FE 830-834 can also have multiple direct links 815 (e.g., secure connection links established through multiple physical links) to another edge FE 830-834, and/or to a datacenter hub FE 836. Again, the different links in some embodiments can use different types of physical links or the same type of physical links. Also, in some embodiments, a first edge FE of a first branch site can connect to a second edge FE of a second branch site (1) directly through one or more links 815, (2) through a cloud gateway or datacenter hub to which the first edge FE connects through two or more links 815, or (3) through another edge FE of another branch site that can augment its role to that of a hub FE. Hence, in some embodiments, a first edge FE (e.g., 834) of a first branch site (e.g., 824) can use multiple SD-WAN links 815 to reach a second edge FE (e.g., 830) of a second branch site (e.g., 820), or a hub FE 836 of a datacenter site 826.

The cloud gateway 805 in some embodiments is used to connect two SD-WAN FEs 830-836 through at least two secure connection links 815 between the gateway 805 and the two FEs at the two SD-WAN sites (e.g., branch sites 820-824 or datacenter site 826). In some embodiments, the cloud gateway 805 also provides network data from one multi-machine site to another multi-machine site (e.g., provides the accessible subnets of one site to another site). Like the cloud gateway 805, the hub FE 836 of the datacenter 826 in some embodiments can be used to connect two SD-WAN FEs 830-834 of two branch sites through at least two secure connection links 815 between the hub 836 and the two FEs at the two branch sites 820-824.

In some embodiments, each secure connection link between two SD-WAN FEs (i.e., CGW 805 and edge FEs 830-836) is formed as a VPN tunnel between the two FEs. In this example, the collection of the SD-WAN FEs (e.g., FEs 830-836 and gateways 805) and the secure connections 815 between the FEs forms the SD-WAN 800 for the particular entity that spans at least the public or private cloud datacenter 810 to connect the branch and datacenter sites 820-826.

In some embodiments, secure connection links are defined between gateways in different public cloud datacenters to allow paths through the virtual network to traverse from one public cloud datacenter to another, while no such links are defined in other embodiments. Also, in some embodiments, the gateway 805 is a multi-tenant gateway that is used to define other virtual networks for other entities (e.g., other companies, organizations, etc.). Some such embodiments use tenant identifiers to create tunnels between a gateway and edge FE of a particular entity, and then use tunnel identifiers of the created tunnels to allow the gateway to differentiate packet flows that it receives from edge FEs of one entity from packet flows that it receives along other tunnels of other entities. In other embodiments, gateways are single-tenant and are specifically deployed to be used by just one entity.

The SD-WAN 800 includes a cluster of controllers 840 that serve as a central point for managing (e.g., defining and modifying) configuration data that is provided to the edge FEs and/or gateways to configure some or all of the operations. In some embodiments, this controller cluster 840 is in one or more public cloud datacenters, while in other embodiments it is in one or more private datacenters. In some embodiments, the controller cluster 840 has a set of manager servers that define and modify the configuration data, and a set of controller servers that distribute the configuration data to the edge FEs, hubs and/or gateways. In some embodiments, the controller cluster 840 directs edge FEs and hubs to use certain gateways (i.e., assigns a gateway to the edge FEs and hubs). In some embodiments, some or all of the controller cluster's functionality is performed by a cloud gateway (e.g., cloud gateway 805). The controller cluster 840 also provides next hop forwarding rules and load balancing criteria in some embodiments.

Many of the above-described features and applications are implemented as software processes that are specified as a set of instructions recorded on a computer-readable storage medium (also referred to as computer-readable medium). When these instructions are executed by one or more processing unit(s) (e.g., one or more processors, cores of processors, or other processing units), they cause the processing unit(s) to perform the actions indicated in the instructions. Examples of computer-readable media include, but are not limited to, CD-ROMs, flash drives, RAM chips, hard drives, EPROMs, etc. The computer-readable media does not include carrier waves and electronic signals passing wirelessly or over wired connections.

In this specification, the term “software” is meant to include firmware residing in read-only memory or applications stored in magnetic storage, which can be read into memory for processing by a processor. Also, in some embodiments, multiple software inventions can be implemented as sub-parts of a larger program while remaining distinct software inventions. In some embodiments, multiple software inventions can also be implemented as separate programs. Finally, any combination of separate programs that together implement a software invention described here is within the scope of the invention. In some embodiments, the software programs, when installed to operate on one or more electronic systems, define one or more specific machine implementations that execute and perform the operations of the software programs.

FIG. 9 conceptually illustrates a computer system 900 with which some embodiments of the invention are implemented. The computer system 900 can be used to implement any of the above-described hosts, controllers, gateway, and edge forwarding elements. As such, it can be used to execute any of the above described processes. This computer system 900 includes various types of non-transitory machine-readable media and interfaces for various other types of machine-readable media. Computer system 900 includes a bus 905, processing unit(s) 910, a system memory 925, a read-only memory 930, a permanent storage device 935, input devices 940, and output devices 945.

The bus 905 collectively represents all system, peripheral, and chipset buses that communicatively connect the numerous internal devices of the computer system 900. For instance, the bus 905 communicatively connects the processing unit(s) 910 with the read-only memory 930, the system memory 925, and the permanent storage device 935.

From these various memory units, the processing unit(s) 910 retrieve instructions to execute and data to process in order to execute the processes of the invention. The processing unit(s) 910 may be a single processor or a multi-core processor in different embodiments. The read-only-memory (ROM) 930 stores static data and instructions that are needed by the processing unit(s) 910 and other modules of the computer system 900. The permanent storage device 935, on the other hand, is a read-and-write memory device. This device 935 is a non-volatile memory unit that stores instructions and data even when the computer system 900 is off. Some embodiments of the invention use a mass-storage device (such as a magnetic or optical disk and its corresponding disk drive) as the permanent storage device 935.

Other embodiments use a removable storage device (such as a floppy disk, flash drive, etc.) as the permanent storage device. Like the permanent storage device 935, the system memory 925 is a read-and-write memory device. However, unlike storage device 935, the system memory 925 is a volatile read-and-write memory, such as random access memory. The system memory 925 stores some of the instructions and data that the processor needs at runtime. In some embodiments, the invention's processes are stored in the system memory 925, the permanent storage device 935, and/or the read-only memory 930. From these various memory units, the processing unit(s) 910 retrieve instructions to execute and data to process in order to execute the processes of some embodiments.

The bus 905 also connects to the input and output devices 940 and 945. The input devices 940 enable the user to communicate information and select commands to the computer system 900. The input devices 940 include alphanumeric keyboards and pointing devices (also called “cursor control devices”). The output devices 945 display images generated by the computer system 900. The output devices 945 include printers and display devices, such as cathode ray tubes (CRT) or liquid crystal displays (LCD). Some embodiments include devices such as touchscreens that function as both input and output devices 940 and 945.

Finally, as shown in FIG. 9, bus 905 also couples computer system 900 to a network 965 through a network adapter (not shown). In this manner, the computer 900 can be a part of a network of computers (such as a local area network (“LAN”), a wide area network (“WAN”), or an Intranet), or a network of networks (such as the Internet). Any or all components of computer system 900 may be used in conjunction with the invention.

Some embodiments include electronic components, such as microprocessors, storage and memory that store computer program instructions in a machine-readable or computer-readable medium (alternatively referred to as computer-readable storage media, machine-readable media, or machine-readable storage media). Some examples of such computer-readable media include RAM, ROM, read-only compact discs (CD-ROM), recordable compact discs (CD-R), rewritable compact discs (CD-RW), read-only digital versatile discs (e.g., DVD-ROM, dual-layer DVD-ROM), a variety of recordable/rewritable DVDs (e.g., DVD-RAM, DVD-RW, DVD+RW, etc.), flash memory (e.g., SD cards, mini-SD cards, micro-SD cards, etc.), magnetic and/or solid state hard drives, read-only and recordable Blu-Ray® discs, ultra-density optical discs, any other optical or magnetic media, and floppy disks. The computer-readable media may store a computer program that is executable by at least one processing unit and includes sets of instructions for performing various operations. Examples of computer programs or computer code include machine code, such as is produced by a compiler, and files including higher-level code that are executed by a computer, an electronic component, or a microprocessor using an interpreter.

While the above discussion primarily refers to microprocessor or multi-core processors that execute software, some embodiments are performed by one or more integrated circuits, such as application-specific integrated circuits (ASICs) or field-programmable gate arrays (FPGAs). In some embodiments, such integrated circuits execute instructions that are stored on the circuit itself.

As used in this specification, the terms “computer”, “server”, “processor”, and “memory” all refer to electronic or other technological devices. These terms exclude people or groups of people. For the purposes of the specification, the terms “display” or “displaying” mean displaying on an electronic device. As used in this specification, the terms “computer-readable medium,” “computer-readable media,” and “machine-readable medium” are entirely restricted to tangible, physical objects that store information in a form that is readable by a computer. These terms exclude any wireless signals, wired download signals, and any other ephemeral or transitory signals.

While the invention has been described with reference to numerous specific details, one of ordinary skill in the art will recognize that the invention can be embodied in other specific forms without departing from the spirit of the invention. In addition, a number of the figures (including FIGS. 2, 5, and 7) conceptually illustrate processes. The specific operations of these processes may not be performed in the exact order shown and described. The specific operations may not be performed in one continuous series of operations, and different specific operations may be performed in different embodiments. Furthermore, the process could be implemented using several sub-processes, or as part of a larger macro process. Thus, one of ordinary skill in the art would understand that the invention is not to be limited by the foregoing illustrative details, but rather is to be defined by the appended claims.

Claims

1. A method of sending data messages from an edge router at a first location of an enterprise network to a Saas (software as a service) application server provided by a third-party at a second location, the method comprising: receiving, from a DNS (domain name system) first server, a resolution for a particular destination network address for the SaaS application server at the second location;from a second server, obtaining an identifier for a first cloud gateway from a plurality of cloud gateways at a plurality of locations through which the particular destination address for the SaaS application server can be reached, the first cloud gateway farther from the first location than a second cloud gateway in the plurality of cloud gateways but closer to the second location than the second cloud gateway; andusing an optimized SD-WAN connection to the first cloud gateway using at least a set of latency scores to forward data messages for the first cloud gateway to the SaaS application at the second location, the set of latency scores being determined by the edge router links between the edge router and the plurality of cloud gateways, the set of latency scores comprising a first latency score for a link between the edge router and the first cloud gateway and a second latency score for a link between the first cloud gateway and the SaaS application server.

2. The method of claim 1, wherein: a first round-trip time (RTT) associated with using the first cloud gateway is longer than a second RTT associated with using the second cloud gateway; andthe first RTT is longer than the second RTT as a result of the first cloud gateway being farther from the first location than the second cloud gateway.

3. The method of claim 1, wherein receiving, from the second server the identifier for the first cloud gateway further comprises receiving, from the second server, location information and the set of latency scores associated with the first cloud gateway.

4. The method of claim 3, wherein the set of latency scores comprises a first latency score associated with a first link between the edge router and the first cloud gateway and a second latency score associated with a second link between the first cloud gateway and the SaaS application server.

5. The method of claim 4, wherein: the set of latency scores is a first set of latency scores;the second cloud gateway is associated with a second set of latency scores comprising a third latency score associated with a third link between the edge router and the second cloud gateway and a fourth latency score associated with a fourth link between the second cloud gateway and the SaaS application server;the first latency score associated with the first link between the edge router and the first cloud gateway and the fourth latency score associated with the fourth link between the second cloud gateway and the SaaS application server comprise high latency scores; andthe second latency score associated with the second link between the first cloud gateway and the SaaS application server and the third latency score associated with the third link between the edge router and the second cloud gateway comprise low latency scores.

6. The method of claim 5, wherein: the SaaS application server is associated with a low latency requirement;the fourth latency score associated with the fourth link between the second cloud gateway and the SaaS application server exceeds a threshold specified by the low latency requirement for the SaaS application server; andthe second latency score associated with the second link between the first cloud gateway and the SaaS application server does not exceed the threshold specified by the low latency requirement for the SaaS application server.

7. The method of claim 1, wherein the second server comprises a management and control server.

8. The method of claim 1, wherein the second server collects metrics from the plurality of cloud gateways, and uses the collected metrics to identify the first cloud gateway as a most optimal cloud gateway for use by the edge router at the first location to forward data messages to the SaaS application server at the second location.

9. The method of claim 8, wherein the SaaS application server is a first SaaS application server, wherein the management and control server provides an identifier for the second cloud gateway for use by the edge router at the first location to send data messages to a second SaaS application server at a third location.

10. The method of claim 1, wherein prior to receiving the identifier for the first cloud gateway, the second cloud gateway is assigned to the edge router as a default cloud gateway for sending data messages to the SaaS application server.

11. A non-transitory machine readable medium storing a program for execution by a set of processing units, the program for sending data messages from an edge router at a first location of an enterprise network to a SaaS (software as a service) application server provided by a third-party at a second location, the program comprising sets of instructions for: receiving, from a DNS (domain name system) first server, a resolution for a particular destination network address for the SaaS application server at the second location;from a second server, obtaining an identifier for a first cloud gateway from a plurality of cloud gateways at a plurality of locations through which the particular destination address for the SaaS application server can be reached, the first cloud gateway farther from the first location than a second cloud gateway in the plurality of cloud gateways but closer to the second location than the second cloud gateway; andusing an optimized SD-WAN connection to the first cloud gateway using at least a set of latency scores to forward data messages for the first cloud gateway to the SaaS application at the second location, the set of latency scores being determined by the edge router links between the edge router and the plurality of cloud gateways, the set of latency scores comprising a first latency score for a link between the edge router and the first cloud gateway and a second latency score for a link between the first cloud gateway and the SaaS application server.

12. The non-transitory machine readable medium of claim 11, wherein: a first round-trip time (RTT) associated with using the first cloud gateway is longer than a second RTT associated with using the second cloud gateway; andthe first RTT is longer than the second RTT as a result of the first cloud gateway being farther from the first location than the second cloud gateway.

13. The non-transitory machine readable medium of claim 11, wherein the set of instructions for receiving, from the second server the identifier for the first cloud gateway further comprises a set of instructions for receiving, from the second server, location information and the set of latency scores associated with the first cloud gateway.

14. The non-transitory machine readable medium of claim 13, wherein the set of latency scores comprises a first latency score associated with a first link between the edge router and the first cloud gateway and a second latency score associated with a second link between the first cloud gateway and the SaaS application server.

15. The non-transitory machine readable medium of claim 14, wherein: the set of latency scores is a first set of latency scores;the second cloud gateway is associated with a second set of latency scores comprising a third latency score associated with a third link between the edge router and the second cloud gateway and a fourth latency score associated with a fourth link between the second cloud gateway and the SaaS application server;the first latency score associated with the first link between the edge router and the first cloud gateway and the fourth latency score associated with the fourth link between the second cloud gateway and the SaaS application server comprise high latency scores; andthe second latency score associated with the second link between the first cloud gateway and the SaaS application server and the third latency score associated with the third link between the edge router and the second cloud gateway comprise low latency scores.

16. The non-transitory machine readable medium of claim 15, wherein: the SaaS application server is associated with a low latency requirement;the fourth latency score associated with the fourth link between the second cloud gateway and the SaaS application server exceeds a threshold specified by the low latency requirement for the SaaS application server; andthe second latency score associated with the second link between the first cloud gateway and the SaaS application server does not exceed the threshold specified by the low latency requirement for the SaaS application server.

17. The non-transitory machine readable medium of claim 11, wherein the second server comprises a management and control server.

18. The non-transitory machine readable medium of claim 11, wherein the second server collects metrics from the plurality of cloud gateways, and uses the collected metrics to identify the first cloud gateway as a most optimal cloud gateway for use by the edge router at the first location to forward data messages to the SaaS application server at the second location.

19. The non-transitory machine readable medium of claim 18, wherein the SaaS application server is a first SaaS application server, wherein the management and control server provides an identifier for the second cloud gateway for use by the edge router at the first location to send data messages to a second SaaS application server at a third location.

20. The non-transitory machine readable medium of claim 11, wherein prior to receiving the identifier for the first cloud gateway, the second cloud gateway is assigned to the edge router as a default cloud gateway for sending data messages to the SaaS application server.