High-speed event processing mechanism in distributed systems

Abstract

A method for use in a distributed system configured to be implemented within one or more clusters of a communication network for transmitting messages within the distributed system, the method and the distributed system being characterized in that the messages include a data payload and a header, and the header includes a tuple of metadata for the data payload.

Description

[Technical Field]

[0001] This disclosure relates generally to the field of communication systems, and more particularly to software architectures for high-scale distributed systems.

[0002] CL: Cluster Leader CM: Cluster Manager EM: Element Manager, an entity that resides in every node of a cluster to manage elements as requested by the CM. FCE: Forwarding Cluster Element FEM: Fabric Element Manager FM: Forwarding Manager FRR: Fast Re-Route RM: Routing Manager WB: White-box Container: A container is an isolated execution environment that operates as a full-featured configuration with its own users, file system, processes, libraries, code, network stack, etc., ensuring that you get identical behavior from the software contained within such a container regardless of the system on which it runs. Control Plane: A logical layer that contains all applications related to all functions and processes that determine the routes used to manage the data plane, etc. This definition includes, but is not limited to, configuration engines, routing stacks, routing protocols, spanning trees, IdPs (identity providers), and user-facing services. Data Plane: The logical layer containing all functions and processes related to the transfer of data packets / frames from one interface to another through the system. This definition includes, but is not limited to, routing through physical ports as well as routing through intermediate backplane ports that allow data packets to reach physical ports that are not immediately reachable. [Background technology]

[0003] In the realm of high-scale distributed systems, one of the challenges in the software architecture domain is the ability to efficiently transmit messages to interested system users.

[0004] Distributed messaging is typically based on the concept of reliable message queues: messages are queued asynchronously between client applications and messaging systems, and such distributed messaging systems offer the benefits of reliability, scalability, and durability.

[0005] However, as application messages become more complex, the overhead caused by classification and forwarding decisions becomes a significant bottleneck, especially as the system itself scales.

[0006] As part of managing such a distributed system consisting of hundreds of nodes, each node typically forwards many messages to its cluster manager (CM). Clearly, such processing involves a huge number of messages arriving at each cluster manager from all nodes contained within each cluster, and therefore, it is necessary to improve the effectiveness of such processing to avoid excessive delays in the transmission of these messages.

[0007] Therefore, the present disclosure aims to solve the problem of effectively transmitting messages in high-scale distributed systems. Summary of the Invention [Problem to be solved by the invention]

[0008] The present disclosure can be summarized by reference to the appended claims.

[0009] It is an object of the present disclosure to provide a novel method and a novel system that allows for efficient forwarding of messages within a system that is a distributed system.

[0010] Another object of the present disclosure is to provide a novel method and system that allows distinguishing important messages from ordinary messages, and consequently allows important messages to be processed sooner than ordinary messages.

[0011] Another object of the present disclosure is to provide a novel method and a novel system that utilizes a high-speed tag-based mechanism for transferring messages within a distributed system.

[0012] Other objects of the present disclosure will become apparent from the following description. [Means for solving the problem]

[0013] According to a first embodiment of the present invention, there is provided a method for use in a distributed system configured to be implemented in one or more clusters of a communication network for transmitting messages within the distributed system, wherein the messages comprise a data payload and a header, and the header comprises a tuple of metadata for the data payload.

[0014] As used herein and in the claims, the term "tuple" is used to denote a finite sequence of conditions, for example, a single column in a relational database.

[0015] As used herein and in the claims, the term "metadata" is used to refer to data that provides information about other data but does not provide the content of the data, such as the text of a message or an image itself. In an embodiment of the solution provided by the present disclosure, the metadata includes at least one member of the group consisting of a target, a source, an event type (i.e., a classification of the event), and event-related data.

[0016] According to another embodiment, the metadata includes at least one member of the group consisting of a target, a source, an event type, and event-related data.

[0017] According to another embodiment, each of said members of at least one of said groups has a binary value, whereby said header is suitable for providing an indication whether the message is an important message or not.

[0018] According to yet another embodiment, each of the one or more clusters includes a cluster manager configured to determine whether a message is an important message based on one or more values ​​of at least one of the members of the group. For example, if one of the members of the group is defined as having a higher weight than the other members, e.g., if the event type value in the header of the message is "1" and all other members of the group have values ​​"0," the message is processed as an important message. Alternatively, if two or more members, e.g., an event-related data member and a target member, each have a value of "1" and the other two members have a value of "0," the system operation definition is configured to consider the message to be an important message. Obviously, if all four members have a value of "1," the message is considered to be an important message.

[0019] In yet other embodiments of the present disclosure, if a message is determined to be an important message, it is transmitted without further processing, e.g., forwarded or relayed as necessary, without further examining the message's headers and without parsing the message to extract the data payload from the message.

[0020] According to yet another aspect of the present disclosure, there is provided a distributed system for use in a communications network, the distributed system including a plurality of nodes arranged in the form of a plurality of clusters and configured to transmit messages within the distributed system, the messages including a data payload and a header, the header including a tuple of metadata associated with the data payload.

[0021] According to another embodiment of an aspect of the present disclosure, the metadata allows for determining how a message associated with the metadata is to be processed.

[0022] In yet another embodiment of this aspect of the present disclosure, the metadata includes at least one member of the group consisting of a target, a source, an event type, and event-related data.

[0023] According to yet another embodiment, each of the members of at least one of the groups has a binary value.

[0024] According to another embodiment, each of the plurality of clusters includes a cluster manager configured to determine whether a message is an important message based on one or more values ​​of at least one of the members of the group.

[0025] In yet another embodiment of this aspect of the present disclosure, the cluster manager is configured to determine whether the message is an important message based on the value of the event type.

[0026] According to yet another embodiment of an aspect of the present disclosure, if the message is determined to be an important message, the message is forwarded without further processing.

[0027] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate several embodiments of the present disclosure and, together with the description, serve to explain the principles of these embodiments disclosed herein. [Brief explanation of the drawings]

[0028] [Figure 1] 1 is a partial diagram of a distributed system in which important messages are separated from ordinary messages, according to an embodiment of the present invention. [Figure 2] 1 is a flow diagram illustrating a method for implementing the inventive solution according to an embodiment of the present invention; DETAILED DESCRIPTION OF THE INVENTION

[0029] Some of the specific details and values ​​in the following detailed description illustrate particular examples of the present disclosure. However, this description is illustrative and is not intended to limit the scope of the present invention. It will be apparent to those skilled in the art that the claimed distributed routing system can also be implemented using other devices known in the art. The scope of the present invention is summarized by reference to the appended claims.

[0030] As part of the management of a distributed system in a communication network consisting of hundreds of nodes, optionally white boxes, each node typically forwards (streams) messages to its cluster manager (CM). Clearly, such processing involves a huge number of messages arriving at each cluster manager, and therefore it is necessary to improve the effectiveness of such processing in order to avoid excessive delays in the processing of the system. The solution provided by the present invention is based on distinguishing important messages from ordinary messages and allows important messages to be processed (e.g., forwarded) earlier than ordinary messages.

[0031] According to an embodiment of the present invention, a novel mechanism for structured binary tag messages to be transferred throughout the system is used.

[0032] This prefix (tag) attached to the message contains metadata for the data payload of the message and includes at least one member of the group consisting of the target to which the message should be forwarded, the source from which the message was generated, the type of event associated with the message, and associated data for the event associated with the message.

[0033] According to an embodiment of the present disclosure, each of the above members included in the metadata simply indicates whether the message is an important message or not. This means, for example, that if one member of the group is defined as having a higher weight than the other members per the system operator's decision, for example, if the event type value in the message's header is "1" and all other members of the group are "0," the message is still processed as an important message. Alternatively, if the system operator decides that two or more members, such as the event-related data member and the target member, each have a value of "1" and the other two members have a value of "0," the system operation definition is configured to consider the message to be an important message. Obviously, if all four members have a value of "1," the message is considered to be an important message.

[0034] Introducing such a structured binary tag as a prefix into the payload of a message allows for quick identification of parameters such as the target to which the message is being delivered, the type of event related to the delivered message, etc. This in turn allows for distinguishing important messages from ordinary messages, and allows for forwarding and / or relaying important messages without the need to deserialize their useful content.

[0035] A message classified as important is immediately transmitted without further analysis or inspection of the message and its data payload. Classification of a message as important or normal is performed by any one or any combination of the values ​​of the metadata prefix components, predefined by system configuration or determined by the system operator.

[0036] 1 is a partial diagram of a distributed system 5 in which important messages are segregated from ordinary messages, according to an embodiment of the present invention. Ordinary messages are forwarded along a "slow path" for non-urgent events (associated with ordinary messages), as opposed to a path for processing messages associated with events marked for relaying (forwarding without further processing).

[0037] The two types of events described above are forwarded to the cluster manager (CM) 10. Information related to these events is forwarded using messages generated by the application, which then sends each message with these metadata tagged events included in the message's header. The messages are sent toward the local node's element manager (EM) 25, which resides in the cluster leader (CL) 30, and the element manager 25 then forwards the messages toward the cluster manager 10 connection. In the cluster manager 10, messages are classified based on the information included in each message's header. If the value included in the metadata indicates that the event is important, the message should be transmitted without further processing. Once the message (and its associated event, respectively) is determined to be an important or ordinary message, the event receiving thread 20 relays the event to a relay event path (for important messages) and then to other nodes 35, or forwards ordinary messages to the internal fabric capacity event path 15, as shown in FIG. 1.

[0038] FIG. 2 shows a flow diagram illustrating a method for implementing the inventive solution according to an embodiment of the present invention.

[0039] In step 200, a message is generated for transmission through a distributed system operating within a communication network, the message including a data payload and a header, and the header including a tuple of metadata (target, source, event type, and event data).

[0040] The message then travels through the distributed system and arrives at the cluster manager (step 210).

[0041] The cluster manager detects the values ​​of the components contained within the header (step 220) and classifies the message as important or normal based on one or more of these values, as determined by the system operator (step 230).

[0042] If the cluster manager determines that the message is important, it does not further process the message and transmits it toward its destination without undue delay (step 240). Optionally, classification capabilities are provided on the part of the cluster manager by examining the event type. Events tagged as having a type that requires immediate relaying are forwarded without further examination and analysis.

[0043] By the illustrated method, event-related messages are forwarded, relayed, and handled more efficiently with greater capacity, i.e., the method results in better resolution of event processing, resulting in more optimal and faster event processing.

[0044] The present invention has been described using detailed descriptions of embodiments that are provided by way of example only and are not intended to limit the scope of the invention. The described embodiments include different configurations, and not all configurations are required in all embodiments of the invention. Some embodiments of the invention utilize only some of the configurations or possible combinations of configurations. Variations of the described embodiments of the invention, as well as embodiments of the invention that include different combinations of the configurations shown in the described embodiments, will be apparent to those skilled in the art. The scope of the invention is limited only by the following claims.

Claims

1. A method for use in a distributed system configured to be implemented in one or more clusters of a communication network for transmitting messages within the distributed system, wherein the messages include a data payload and a header, and the header includes a tuple of metadata for the data payload.

2. 10. The method of claim 1, wherein the metadata enables a determination of how to process a message associated with the metadata.

3. 10. The method of claim 1, wherein the metadata includes at least one member of the group consisting of a target, a source, an event type, and event-related data.

4. 4. The method of claim 3, wherein each of the members of at least one of the groups has a binary value.

5. 4. The method of claim 3, wherein each of the one or more clusters includes a cluster manager configured to determine whether a message is an important message based on one or more values ​​of the at least one member of the group.

6. 6. The method of claim 5, wherein the cluster manager is configured to determine whether a message is an important message based on the value of the event type.

7. 6. The method of claim 5, wherein if the message is determined to be an important message, the message is forwarded without further processing.

8. A distributed system for use within a communications network, the distributed system comprising a plurality of nodes arranged in the form of a plurality of clusters and configured to transmit messages within the distributed system, the messages comprising a data payload and a header, the header comprising a tuple of metadata associated with the data payload.

9. 9. The distributed system of claim 8, wherein the metadata enables determining how to process a message associated with the metadata.

10. 10. The distributed system of claim 8, wherein the metadata includes at least one member of the group consisting of a target, a source, an event type, and event-related data.

11. 11. The distributed system of claim 10, wherein each of the members of at least one of the groups has a binary value.

12. 11. The distributed system of claim 10, wherein each of the plurality of clusters includes a cluster manager configured to determine whether a message is an important message based on one or more values ​​of the at least one member of the group.

13. 13. The distributed system of claim 12, wherein the cluster manager is configured to determine whether a message is an important message based on the value of the event type.

14. 13. The distributed system of claim 12, wherein if the message is determined to be an important message, the message is forwarded without further processing.

Images