Browse Prior Art Database

MECHANISM FOR THE EXTRACTION OF MULTIPLE LARGE BINARY DATA OBJECTS WITHIN A DATA STREAM OF UNLIMITED SIZE

IP.com Disclosure Number: IPCOM000013646D
Original Publication Date: 2001-Sep-01
Included in the Prior Art Database: 2003-Jun-18
Document File: 2 page(s) / 42K

Publishing Venue

IBM

Abstract

Problem: Communications datastreams often contain large objects (ie contiguous streams of binary bits which often represent multimedia entities; eg images, sound, video). These objects must be removed from the data stream without error and processed while the datastream is still being received. An example of one such datastream is the DRDA protocol used to move data between the clients and servers of relational databases. The problem is: How do you continuously process normal command or request information from a data stream, and simultaneously remove the large binary objects from the data stream as well?

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 2

  MECHANISM FOR THE EXTRACTION OF MULTIPLE LARGE BINARY DATA OBJECTS WITHIN A DATA STREAM OF UNLIMITED SIZE

Problem: Communications datastreams often contain large objects (ie contiguous streams of binary bits which often represent multimedia entities; eg images, sound, video). These objects must be removed from the data stream without error and processed while the datastream is still being received. An example of one such datastream is the DRDA protocol used to move data between the clients and servers of relational databases. The problem is: How do you continuously process normal command or request information from a data stream, and simultaneously remove the large binary objects from the data stream as well?

Assumptions: It is assumed that the datastream contains multiple objects of a finite size, and that those objects contain length information identifying the overall length of the object. Furthermore the datastream is itself divided into segments of finite size and that these segments contain length information identifying the length of the segment. Furthermore it is assumed that sets of the large objects (LOBs) are interspersed with descriptive objects (requests). The requests identify the number and sizes of the LOBs which follow in the datastream.

Abstract: The technique of dynamically extracting large objects from the described datastream is comprised of a combination of 4 features:
a) Double buffering.
b) Triple object length verification.
c) Double parsing (static and dynamic).
d) Request to Large Object Correlation information.

Advantages: The use of this technique allows large objects and request to be extracted continuously from a data stream without the need for collecting a large object in storage or memory buffers. The technique assures integrity of the data objects and assures that the flow-control integrity of the half-duplex protocol is not compromised. It furthermore allows requests and large objects to be interspersed in the datastream. Without this technique, either the entire datastream must be homogenous (no mixed requests and large objects, or the objects and requests must be sent separately) or the entire message must be collected into either memory buffers or files.

The following rules must be adhered to:

Every set of large objects in the datastream must be preceded by a request object which identifies the number and size of large

objects to follow. Every object must begin with a header field which identifies the length of that object, a request id number for that object, and a field

which indicates whether another object follows the given object in the datastream. The request id of a large object must match the request id of any preceding object.

The request id of a requester object must be different from the request id of a preceding large object.

No large obj...