System and method for storing large messages

Abstract

A large message can be stored by separating the message into an envelope portion containing information such as headers, protocols, and addresses, and a payload portion containing items such as file attachments. The envelope portion can be stored in local storage, while the payload can be stored to a persistent store. The message can be processed incrementally, such that the entire message is never in system memory. Once the envelope portion is processed, the payload portion can be read in increments without being processed, and those increments written directly to the persistent store. Alternatively, the payload can be streamed to the persistent store. A pointer in the envelope can then be used to locate and retrieve attachments from persistent storage.

Description

COPYRIGHT NOTICE

A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document of the patent disclosure, as it appears in the Patent and Trademark Office patent file or records, but otherwise reserves all copyright rights whatsoever.

CROSS-REFERENCED CASES

The following applications are cross-referenced and incorporated herein by reference:

U.S. patent application Ser. No. 10/404,552 filed Apr. 1, 2003, to Mike Blevins et al. and entitled, “COLLABORATIVE BUSINESS PLUG-IN FRAMEWORK”;

U.S. patent application Ser. No. 10/404,684 filed Apr. 1, 2003, to Mike Blevins et al. and entitled, “SYSTEMS AND METHODS FOR BUSINESS PROCESS PLUG-IN DEVELOPMENT”; and

U.S. patent application Ser. No. 10/404,666 filed Apr. 1, 2003, to David Wiser et al. and entitled “Single Servlets for B2B Message Routing.”

FIELD OF THE INVENTION

The present invention relates to the storage of large messages in a computer system or on a computer network.

BACKGROUND

Existing integration and messaging systems have problems handling large messages. Incoming messages are read into memory in their entirety, such that when a number of large messages are received a system can crash due to a lack of available memory. Some systems try to prevent these problems by limiting the size of messages that can be processed through a system, but this approach is undesirable to users needing to send messages that may occasionally exceed that limitation.

Another existing approach utilizes in-database persistence and in-memory caching on a hub. Persistence saves enough data for recovery purposes, and caching allows messages to be serialized to a Java Message Service (JMS). This allows JMS to enqueue faster, and allows a JMS dequeue to request the message from a cache without having to redo expensive operations like deserialization, decryption, and XML parsing. The problem still exists in that it is necessary to read an entire message into memory in order to process the message.

BRIEF SUMMARY

Systems and methods in accordance with embodiments of the present invention can overcome deficiencies in existing messaging systems by changing the way in which messages are processed and stored. An integration component can receive an incoming message, such as from a Web server. The integration component can separate the message into an “envelope” portion, which can contain information such as headers, protocols, and addresses, and a “payload” portion, which can contain items such as file attachments. The integration component can write the envelope portion to local memory, and can write the payload portion to at least one persistent store. A pointer can be placed in the envelope to identify the location of the payload in the persistent store. Applications can then use the envelope to locate the payload in a persistent store.

An integration component can also process a message incrementally. The integration component can process portions of the message until the payload portion is reached. The integration component can then stop processing the message, but can continue to read the message in increments and write those increments to a persistent store. Parsers such as MIME parsers and XML parsers can be used by the integration component to process the message. Alternatively, the integration component can process the message as a stream, or at least write the payload portion to the persistent store as a stream.

Other features, aspects, and objects of the invention can be obtained from a review of the specification, the figures, and the claims.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a diagram of a system in accordance with one embodiment of the present invention.

FIG. 2 is a diagram of a message that can be processed using the system of FIG. 1.

DETAILED DESCRIPTION

In systems and methods in accordance with embodiments of the present invention, “large” messages, such as large business messages in XML format, an be processed in a Web server or integration application. These business messages can be Java Message Service (JMS) messages, for example, which can utilize distributed destinations in a cluster. A large business message can be any message that may have an attachment or a large amount of text, for example, which can have an overall message size of at or above 1 MB, at or above 10 MB, at or above 50 MB, or even at or above 100 MB. For example, company A can send a message to company B that has a file size of 100 MB. The integration system receiving that message will have to process and resend the entire message. In existing systems, it is necessary to read the entire message into memory before writing the message to disk. The read and write are each done in one complete step. Present systems also have to parse the entire message.

As shown in FIG. 1, when a message from a company A 100 is first received to an integration component 104 from a Web server 102, the message can be read into local memory 106. Local memory can be any appropriate storage medium, such as may be located on the Web server itself, in a cluster containing the Web server, or on a network node accessible to the Web server. If several large messages are received by the Web server 102, the server may eventually run out of memory. In a system in accordance with one embodiment of the present invention, portions of the body of each message can be stored in persistent storage 108 instead of being completely stored in local memory 106. There are at least two types of persistent storage, including file-based persistence stores and data-based persistence stores.

Continuing with the example, company B 110 can be working with an integration application. When the message arrives at the Web server 102 for company B 110, the message can arrive on a socket on the network. Portions of the message can be stored somewhat directly to the persistent message store 108 instead of being read entirely into local memory 106. One way to do this is to read the message in increments, or small portions, and write those small portions to storage. For example, the 100 MB message could have a 4 MB portion read into local memory 106, then have that 4 MB portion written to persistent storage 108. Then another 4 MB portion could be read into local memory and written to persistent storage. This process could continue until the entire body portion of the message is in persistent storage 108. Although portions of the entire message may be in memory at one point or another, there would only be up to 4 MB of the message in local memory at any given time. The user can configure the persistent store 108 so that the message is sent to a file or to a database, for example. The portion size can be any size appropriate for the size of the message or the capacity of the system, such as portions of 1 MB, 5 MB, or 10 MB. The portion size can also be a percentage of the overall file size, such as 1%, 5%, 10%, or 25%, for example.

When a message is processed using an integration application or integration server, for example, the message can use a storage method referred to herein as “envelope plus payload.” The message can be processed in the server to separate the contents to be placed in the “envelope” from contents to be placed in the “payload.” This is shown, for example, in the diagram of FIG. 2. Headers 202, 204 of a message 200 can be extracted by an integration server, as the headers may be all the server requires to process the message 200. A header can identify the protocol under which the message is sent, such as an XOCP protocol. The protocol can be used to help identify the headers 202, 204 and the body 206 of the message. It can be important in certain systems to identify the message protocol, as protocols such as RosettaNet and ebXML have different packaging semantics than a protocol such as XOCP. The headers can be placed in the envelope 214, which can be stored in local memory. The body 206 of the message 200, which can contain several attachments 208, 210, 212, for example, can be placed into the payload 218. The payload can be stored in persistent storage on the server, in the cluster, or on the network. The envelope 214 can contain a pointer 216 to the location of the payload 218.

Since a message can contain a body with multiple parts, the payload can be designed to contain multiple parts as well. While processing a message in the server, however, only the envelope may be needed. The payload can belong to the user of a B2B server, for example, or an application riding on top of an integration server. The payload can be stored to persistent storage, so that the full payload is never stored in memory. A server or any application can simply deal with the envelope, which can contain pointers to the payload. When an application wants to access any portion of the message, the application can view information contained in the envelope, which can include identification information for the payload parts.

An application can use any pointers in an envelope to extract portions of the body of the message stored in the payload. As the application can retrieve the data from this persistence store, it is not necessary to accumulate everything in local memory on the integration server. A message envelope can contain a pointer to the body of the message, whether there is a single message body or a number of portions, or can contain a pointer for each portion of the body in persistent storage. The number of portions can include a number of attachments, for example. It is not necessary for the integration system to process the attachments to a message, so the system can simply write the attachments to storage, either all together in one block of memory or individually. The pointer can point to the location at which a portion of the message body begins in memory, or can point to the boundaries of a given body portion in memory, for example.

An envelope can contain other useful information about a message, such as the address of the sender and/or the address of the recipient. Each of these addresses can each be a URL, for example. The envelope can also contain the protocol of the message and possibly the protocol of any body portion, if applicable. The envelope can contain message text. The envelope can also contain information about each attachment in the body, such as title, file type, and historical information.

At least two levels of parsing can be used to process a message. A low-level parsing mechanism can be used to decode transfer protocols such as MIME or UUENCODE. The low-level parser can receive the byte stream and identify the parts of the message, such as a text portion and a binary attachment. A second level of parsing, such as XML parsing, can be used to read headers and body portions, which can be in XML or another appropriate messaging or mark-up language.

A message can arrive from the Internet, for example, and can pass through the Web server into an integration transport layer. First, the message can pass through a MIME parser. Second, the message can be decoded using a second processing layer to determine the appropriate business protocol. The envelope can be created in this transport layer. In the decoding process, which can use the XML parser, the envelope can be filled with headers and other appropriate information. After the headers, a pointer can be placed in the envelope and the MIME parser can stop parsing the message. The MIME parser can know to stop parsing when it hits attachments, for example.

The remainder of the message, which can include at least a portion of the body and any attachments, can then be written directly to persistent storage, either in small data “dumps” or on a data stream. Once the entire message is processed, such that the envelope and payload are created, an application can determine where the message portions reside using pointers in the message envelope. The envelope can be thought of as an “abstract” of the message. Once a user or application gets this abstract, that user or application can extract any portion of the message that is needed. For instance, if there are three attachments, the user or application can choose to extract one or two of the attachments from the persistent storage. When the user deletes the message, the envelope can be used, such as by an integration server or B2B server, to delete the associated portions in the persistent storage.

The foregoing description of preferred embodiments of the present invention has been provided for the purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Many modifications and variations will be apparent to one of ordinary skill in the art. The embodiments were chosen and described in order to best explain the principles of the invention and its practical application, thereby enabling others skilled in the art to understand the invention for various embodiments and with various modifications that are suited to the particular use contemplated. It is intended that the scope of the invention be defined by the following claims and their equivalence.

Claims

1. A system for providing access to a message in a computer environment, comprising: a computer, that includes a memory for in-memory storage of message information;a persistent store for persistent storage of message data; andan integration component that receives a message from a sending application, and then creates both an envelope, and a separate payload, for the message, including reading the message in incremental portions smaller than the total message size, and successively writing those portions to the persistent storage as the payload,stores the payload in the persistent store as a plurality of selectively retrievable portions,creates within the envelope a header information, and pointers to different portions of the payload,stores the envelope in the memory, while the separate payload is stored at the persistent store,provides in-memory access to the envelope by one or more receiving applications, andenables the receiving applications to access the envelope in memory, and to use the pointers in the message envelope to select and retrieve selected portions of the message payload from the persistent store for use by those receiving applications.

2. The system of claim 1, wherein the step of creating a separate envelope and payload includes parsing the message with a protocol parser to identify portions of the message to the integration component.

3. The system of claim 1, wherein the step of creating a separate envelope and payload includes parsing the message with an XML parser to read header and body portions of the message, and to separate the head and body portions into the envelope and payload respectively.

4. The system of claim 1, wherein the step of creating a separate envelope and payload includes processing the message in increments.

5. The system of claim 1 wherein the persistent store is located at the computer that receives the message.

6. The system of claim 1 wherein the persistent store is located at a computer different from the one that receives the message, and wherein the persistent store is shared among a plurality of computers for storage of and access to message payloads for the plurality of computers.

7. A method for providing access to a message in a computer environment, comprising the steps of: providing a computer that includes a memory for in-memory storage of message information,a persistent store for persistent storage of message data, andan integration component;receiving a message from a sending application, at the integration component at the computer;creating both an envelope, and a separate payload, for the message, including reading the message in incremental portions smaller than the total message size, and successively writing those portions to the persistent storage as the payload;storing the payload in the persistent store as a plurality of selectively retrievable portions;creating within the envelope a header information, and pointers to different portions of the payload;storing the envelope in the computer's memory, while the separate payload is stored at the persistent store; andproviding in-memory access to the envelope by one or more receiving applications to access the envelope in memory, and to use the pointers in the message envelope to select and retrieve selected portions of the message payload, from the persistent store, for use by those receiving applications.

8. The method of claim 7, wherein the step of creating a separate envelope and payload includes parsing the message with a protocol parser to identify portions of the message to the integration component.

9. The method of claim 7, wherein the step of creating a separate envelope and payload includes parsing the message with an XML parser to read header and body portions of the message, and to separate the head and body portions into the envelope and payload respectively.

10. The method of claim 7, wherein the step of creating a separate envelope and payload includes processing the message in increments.

11. The method of claim 7 wherein the persistent store is located at the computer that receives the message.

12. The method of claim 7 wherein the persistent store is located at a computer different from the one that receives the message, and wherein the persistent store is shared among a plurality of computers for storage of and access to message payloads for the plurality of computers.

13. A computer readable medium including instructions stored thereon which when executed cause the computer to perform the steps of: receiving at a computer, including a memory for in-memory storage of message information, and a persistent store for persistent storage of message data, a message, from a sending application, at an integration component at the computer;creating both an envelope and a separate payload for the message, including reading the message in incremental portions smaller than the total message size, and successively writing those portions to the persistent storage as the payload;storing the payload in the persistent store as a plurality of selectively retrievable portions;creating within the envelope a header information, and pointers to different portions of the payload;storing the envelope in the computer's memory, while the separate Payload is stored at the persistent store; andproviding in-memory access to the envelope, by one or more receiving applications to access the envelope in memory, and to use the pointers in the message envelope to selectively retrieve selected portions of the message payload from the persistent store, for use by those receiving applications.

14. The computer readable medium of claim 13, wherein the step of creating a separate envelope and payload includes parsing the message with a protocol parser to identify portions of the message to the integration component.

15. The computer readable medium of claim 13, wherein the step of creating a separate envelope and payload includes parsing the message with an XML parser to read header and body portions of the message, and to separate the head and body portions into the envelope and payload respectively.

16. The computer readable medium of claim 13, wherein the step of creating a separate envelope and payload includes processing the message in increments.

17. The computer readable medium of claim 13 wherein the persistent store is located at the computer that receives the message.

18. The computer readable medium of claim 13 wherein the persistent store is located at a computer different from the one that receives the message, and wherein the persistent store is shared among a plurality of computers for storage of and access to message payloads for the plurality of computers.