Abstract: Some embodiments override network or router level path selection with application or server controlled path selection by repurposing the type-of-service (ToS) or differentiated services header field. A mapping table maps different ToS values to different available transit provider paths to a particular destination. A server generating a packet to the destination selects one of the available paths according to any of load balanced, failover, or performance optimization criteria. The server sets the packet header ToS field with the value assigned to the selected path. A router operating in the same network as the server is configured with policy based routing rules that similarly map the ToS values to different transit provider paths to the particular destination network. Upon receiving the server generated packet, the router routes the packet to the destination network through the transit provider path identified in the packet header by the server set ToS value.

Abstract: A test network is provided to test updates to configurations and resources of a distributed platform and to warm servers prior to their deployment in the distributed platform. The test network tests and warms using real-time production traffic of the distributed platform in a manner that does not impact users or performance of the distributed platform. At least one distributed platform caching server passes content requests that it receives to the test network using a connectionless protocol. The test network includes a test server that is loaded with any of a configuration or resource under test or whose cache is to be loaded prior to the server's deployment into the distributed platform. The test network also includes a replay server that receives the requests from the caching server, distributes the requests to the test server, and monitors the test server responses.

Abstract: Some embodiments provide a content delivery network (CDN) solution that affords the CDN control over those elements of customer content that are delivered by third parties. The CDN integrates a distributed set of monitoring agents. Each monitoring agent monitors the delivery performance of third parties to the region in which the agent operates. The CDN uses the performance monitoring information to dynamically manage the content tags to the third-party delivered elements of CDN-customer content. Specifically, a CDN server retrieves the parent page for requested CDN-customer content. The CDN server identifies the region from where the request originates and retrieves the logs from the monitoring agents monitoring from that region. The CDN server then modifies the base page by dynamically removing the tags to the third-party delivered elements that are reported in the monitoring agent logs as being unavailable, inaccessible, or underperforming in the identified region.

Abstract: Some embodiments provide instantaneous and non-blocking content purging across storage servers of a distributed platform. When a server receives a purge operation, it extracts an identifier from the purge operation. The server then generates a content purge pattern from the identifier and injects the pattern to its configuration. Instantaneous purging is then realized as the server averts access to any cached content identified by the pattern. The purging also occurs in a non-blocking fashion as the physical purge of the content occurs in-line with the server's cache miss operation. The content purge pattern causes the server to respond to a subsequently received content request with a cache miss, whereby the server retrieves the requested content from an origin source, serves the retrieved content to the requesting user, and replaces a previously cached copy of the content that is to be purged with the newly retrieved copy.

Abstract: Some embodiments provide instantaneous and non-blocking content purging across storage servers of a distributed platform. When a server receives a purge operation, it extracts an identifier from the purge operation. The server then generates a content purge pattern from the identifier and injects the pattern to its configuration. Instantaneous purging is then realized as the server averts access to any cached content identified by the pattern. The purging also occurs in a non-blocking fashion as the physical purge of the content occurs in-line with the server's cache miss operation. The content purge pattern causes the server to respond to a subsequently received content request with a cache miss, whereby the server retrieves the requested content from an origin source, serves the retrieved content to the requesting user, and replaces a previously cached copy of the content that is to be purged with the newly retrieved copy.

Abstract: Some embodiments increase throughput across a connection between a host and a client by initializing the congestion window for that connection dynamically using a previously settled value from a prior instance of the connection established between the same or similar endpoints. An initialization agent tracks congestion window values for previously established connections between a host and various clients. For the tracked congestion window values of each monitored connection, the initialization agent stores an address identifying the client endpoint. When establishing a new connection, the initialization agent determines if the new connection is a recurring connection. A new connection is recurring when the new connection client address is similar or related to an address identified for a previous monitored connection.

Abstract: Some embodiments provide a proprietary 64-bit consistent distribution scheme that preserves the efficiencies of CARP while providing a significantly more balanced distribution of requests that is on par with schemes reliant on computationally expensive cryptographic hashes. The scheme performs hashing of requested URLs and identifiers of available servers over a 64-bit space while optimizing the hashing to remove computationally expensive operations. Some embodiments provide a variant of the scheme to provide a differentiated distribution on the basis of one or more differentiating factors. A first variant utilizes load factor values to adjust the resulting hashes and to produce a first distribution of differentiated content that varies from a second distribution of undifferentiated content.

Abstract: Some embodiments provide systems and methods for sandboxing content optimization to occur entirely within a network edge or PoP of a CDN. Some embodiments pass a first request for a first URL to a first back-end at the network edge that is configured to cache an optimized instance of the particular object. When the optimized instance of the particular object is not cached at the first back-end, a second request is issued for a second URL identifying a non-optimized instance of the particular object. The second request resolves internally within the network edge to a second back-end that is configured to cache the non-optimized object. The non-optimized object from the second back-end is optimized and passed to the first back-end. The first back-end caches the optimized instance of the non-optimized object and serves the optimized instance to a requesting end user.

Abstract: Some embodiments provide a multi-tenant over-the-top multicast solution that integrates the per user stream customizability of unicast with the large scale streaming efficiencies of multicast. The solution involves an application, different multicast groups streaming an event with different customizations, and a manifest file or metadata identifying the different groups and customizations. The solution leverages the different multicast groups in order to provide different time shifts in the event stream, different quality level encodings of the event stream, and different secondary content to be included with a primary content stream. The application configured with the manifest file or metadata dynamically switches between the groups in order to customize the experience for a user or user device on which the application executes. Switching from multicast to unicast is also supported to supplement available customizations and for failover.

Abstract: Some embodiments provide a proprietary 64-bit consistent distribution scheme that preserves the efficiencies of CARP while providing a significantly more balanced distribution of requests that is on par with schemes reliant on computationally expensive cryptographic hashes. The scheme performs hashing of requested URLs and identifiers of available servers over a 64-bit space while optimizing the hashing to remove computationally expensive operations. Some embodiments provide a variant of the scheme to provide a differentiated distribution on the basis of one or more differentiating factors. A first variant utilizes load factor values to adjust the resulting hashes and to produce a first distribution of differentiated content that varies from a second distribution of undifferentiated content.

Abstract: A test network is provided to test updates to configurations and resources of a distributed platform and to warm servers prior to their deployment in the distributed platform. The test network tests and warms using real-time production traffic of the distributed platform in a manner that does not impact users or performance of the distributed platform. At least one distributed platform caching server passes content requests that it receives to the test network using a connectionless protocol. The test network includes a test server that is loaded with any of a configuration or resource under test or whose cache is to be loaded prior to the server's deployment into the distributed platform. The test network also includes a replay server that receives the requests from the caching server, distributes the requests to the test server, and monitors the test server responses.

Abstract: Some embodiments provide a file system for purging content based on a single traversal of the file system that identifies the directory containing the content without performing a secondary traversal within the directory to target the operation to only the file that are associated with content such that other files contained in the directory are unaffected. The file system supplements traditional directory structures with file-level directories. Each file-level directory is created to contain a root file associated with particular content, different variants of the particular content, and supporting files. Consequently, the file system can complete an operation targeting particular content by performing that operation on the file-level directory containing the particular content, thereby eliminating the need to conduct a file-by-file traversal of the containing directory as a prerequisite to identifying the files associated with the particular content and performing the operation on the files individually.

Abstract: Some embodiments provide systems and methods for determining a server of a distributed hosting system to optimally distribute content to an end user. The method includes identifying an IP address of the end user. Based on the IP address, a set of servers send packets to the end user to derive performance metrics. The performance metrics are used to determine a server from the set of servers that optimally distributes content to the end user. The method modifies a configuration for resolving end user requests such that the optimal server is identified to the end user when the end user requests content from the hosting system. Some embodiments determine the optimal server by providing downloadable content that is embedded with a monitoring tool. The monitoring tool causes the end user to derive performance metrics for the hosting system when downloading a particular object from a set of servers.

Abstract: Some embodiments provide a file system for purging content based on a single traversal of the file system that identifies the directory containing the content without performing a secondary traversal within the directory to target the operation to only the file that are associated with content such that other files contained in the directory are unaffected. The file system supplements traditional directory structures with file-level directories. Each file-level directory is created to contain a root file associated with particular content, different variants of the particular content, and supporting files. Consequently, the file system can complete an operation targeting particular content by performing that operation on the file-level directory containing the particular content, thereby eliminating the need to conduct a file-by-file traversal of the containing directory as a prerequisite to identifying the files associated with the particular content and performing the operation on the files individually.

Abstract: Some embodiments provide a content delivery network (CDN) solution that affords the CDN control over those elements of customer content that are delivered by third parties. The CDN integrates a distributed set of monitoring agents. Each monitoring agent monitors the delivery performance of third parties to the region in which the agent operates. The CDN uses the performance monitoring information to dynamically manage the content tags to the third-party delivered elements of CDN-customer content. Specifically, a CDN server retrieves the parent page for requested CDN-customer content. The CDN server identifies the region from where the request originates and retrieves the logs from the monitoring agents monitoring from that region. The CDN server then modifies the base page by dynamically removing the tags to the third-party delivered elements that are reported in the monitoring agent logs as being unavailable, inaccessible, or underperforming in the identified region.

Abstract: Some embodiments provide a content delivery network (CDN) solution that affords the CDN control over those elements of customer content that are delivered by third parties. The CDN integrates a distributed set of monitoring agents. Each monitoring agent monitors the delivery performance of third parties to the region in which the agent operates. The CDN uses the performance monitoring information to dynamically manage the content tags to the third-party delivered elements of CDN-customer content. Specifically, a CDN server retrieves the parent page for requested CDN-customer content. The CDN server identifies the region from where the request originates and retrieves the logs from the monitoring agents monitoring from that region. The CDN server then modifies the base page by dynamically removing the tags to the third-party delivered elements that are reported in the monitoring agent logs as being unavailable, inaccessible, or underperforming in the identified region.

Abstract: Some embodiments provide a repository that manages configurations for a distributed platform and that automatedly configures servers of the distributed platform with different hierarchical sets of configurations while ensuring integrity and consistency across the servers and in the repository. In some embodiment, the repository includes a data store that stores configurations for a first set of servers that are operated by a first service provider and a second set of servers that are operated by a second service provider. The data store also identifies different sets of configurations to deploy to different sets of servers from the first and second sets of servers. The repository also includes a function processor to automatedly deploy the different sets of configurations to the different sets of servers and to perform functions for updating the configurations in a manner that ensures integrity and consistency.

Abstract: Some embodiments provide instantaneous and non-blocking content purging across storage servers of a distributed platform. When a server receives a purge operation, it extracts an identifier from the purge operation. The server then generates a content purge pattern from the identifier and injects the pattern to its configuration. Instantaneous purging is then realized as the server averts access to any cached content identified by the pattern. The purging also occurs in a non-blocking fashion as the physical purge of the content occurs in-line with the server's cache miss operation. The content purge pattern causes the server to respond to a subsequently received content request with a cache miss, whereby the server retrieves the requested content from an origin source, serves the retrieved content to the requesting user, and replaces a previously cached copy of the content that is to be purged with the newly retrieved copy.

Abstract: Some embodiments provide methods and systems for improving the scalability and redundancy of a distributed content streaming system. Such scalability and redundancy is provided with zero configuration changes to the addressing used by content providers to publish content and zero configuration changes to existing servers of the system. The system includes ingest servers and edge servers. Content providers supply content streams to the ingest servers using a virtual or load balanced address that distributes the content streams across the ingest servers. Accordingly, ingest servers can be added or removed without changing content provider configurations. The ingest servers are configured to notify the edge servers of which content streams are available for streaming at which ingest server. When an ingest server is added to the system, its functionality may be assimilated without modifying the configurations of the other servers. Some embodiments also provide multiple caching layers.

Abstract: Some embodiments provide a repository that manages configurations for a distributed platform and that automatedly configures servers of the distributed platform with different hierarchical sets of configurations while ensuring integrity and consistency across the servers and in the repository. In some embodiment, the repository includes a data store that stores configurations for a first set of servers that are operated by a first service provider and a second set of servers that are operated by a second service provider. The data store also identifies different sets of configurations to deploy to different sets of servers from the first and second sets of servers. The repository also includes a function processor to automatedly deploy the different sets of configurations to the different sets of servers and to perform functions for updating the configurations in a manner that ensures integrity and consistency.