Electrical computers and digital processing systems: memory – Address formation – Hashing
Reexamination Certificate
1999-08-13
2001-11-27
Verbrugge, Kevin (Department: 2185)
Electrical computers and digital processing systems: memory
Address formation
Hashing
Reexamination Certificate
active
06324637
ABSTRACT:
BRIEF DESCRIPTION OF THE INVENTION
This invention relates generally to accessing software objects in an object oriented software environment. More particularly, this invention relates to a technique for reducing the primary memory footprint associated with the process of accessing software objects.
BACKGROUND OF THE INVENTION
FIG. 1
illustrates a computer
20
constructed in accordance with the prior art. The computer
20
includes a central processing unit
22
, which communicates with a set of input/output devices (e.g., keyboard, mouse, video monitor, printer, and the like)
24
over a system bus
26
. Also connected to the system bus is a primary memory (e.g., volatile memory, such as a RAM)
28
and a secondary memory (e.g., mass storage, such as a disk)
30
. The secondary memory stores a set of software objects
32
. The software objects
32
are characterized by an object list
34
, which includes header entries
36
and a central directory
38
, which provides a short-hand characterization of the header entries
36
.
An application program
40
stored in primary memory
28
is executed by the central processing unit
22
. The application program
40
invokes objects
32
to perform specified functions. In particular, the application program
40
accesses a primary memory object index
42
, which basically corresponds to the object list central directory
38
, as discussed below. The primary memory object index
42
allows the application program
40
to identify and load invoked objects
32
stored in secondary memory
30
. After being invoked, the selected objects
44
are resident in primary memory
28
.
FIG. 2
illustrates an object list
34
in accordance with the prior art. The object list
34
includes a set of object list entries
36
A-
36
N. Each entry includes an object name
50
, a header
52
, and data
54
. The header
52
includes information, such as object version, general purpose bit flags, compression method information, a cyclic redundancy check of uncompressed data, compressed data size, uncompressed data size, filename length, and the like. The data
54
includes the instructions or other information associated with the object. Since each entry
36
can be relatively large, a central directory
38
is associated with the object list
34
. The central directory includes a set of central directory entries
48
A-
48
N corresponding to the object list entries
36
A-
36
N. Each central directory entry
48
includes an object name
50
and a pointer
60
to the location of the corresponding object list entry
36
.
As previously indicated, the primary memory object index
42
is typically implemented as the central directory
38
of the object list
34
. The problem with this approach is that the central directory
38
can be relatively large because it includes such information as the object name, the object's location, and miscellaneous information about the object. It is well known that the performance of a computer is largely contingent upon its efficient use of primary memory. An oversized primary memory object index can lead to inferior computer performance.
The foregoing problem is more fully appreciated in connection with a specific example. JAVA™ is a well known computer language developed and licensed by Sun Microsystems, Inc., Mountain View, Calif., the assignee of the present invention. An application program
40
written in JAVA™ is supported by a large object list. The central directory
38
of the object list
34
is scanned through a bootclasspath or application classpath, resulting in the construction of the primary memory object index
42
. The resultant primary memory object index
42
is relatively large.
The JAVA™ Development Kit (JDK™) is an object list
34
which contains the software and tools that developers need to compile, debug, and run applets and applications written using the JAVA™ programming language. Java Archive (JAR) is a platform-independent file format that aggregates many files into one. Multiple JAVA™ applets and their requisite components (e.g., class files, images and sounds) can be bundled in a JAR file and subsequently downloaded to a web browser in a single Hypertext Transport Protocol (HTTP) transaction, greatly improving the download speed. The utility “java.util.jar” provides classes for reading and writing the JAR file format. The JAR file format is based on a “ZIP” file format that is used in JAVA™. The utility “java.util.zip” provides classes for reading and writing the standard ZIP and GZIP file formats. Some of the classes associated with “java.util.zip” that are used by “java.util.jar” include: DeflaterOutputStream, which is a class to implement an output stream filter for compressing data in the “deflate” compression format; InflaterinputStream, which is a class to implement a stream filter for uncompressing data in the “deflate” compression format; ZipEntry, which is a class to represent a ZIP file entry, ZipFile, which is a class to read entries from a zip file; and ZipException, which is a class to handle exceptions. Some of the classes that are used by “java.util.zip” include: checksum, which is an interface representing a data checksum; CRC32, which is a class that can be used to compute the thirty-two bit cyclic redundancy check of a data stream; and ZipEntry, which is a class used to represent a ZIP file entry.
In JDK 1.2, for each ZIP or JAR file on the bootclasspath or application classpath, the file's central directory
38
is scanned and a primary memory object index
42
is built. In the prior art, all of the information from the central directory
38
is loaded into primary memory
28
. This includes the name of the ZIP entry, it's offset in the ZIP file, and about 24 bytes of additional information. This results in about 4,500 ZIP entries in a primary memory object index
42
that takes up approximately 424 KB. This relatively large primary memory footprint is expected to grow as additional classes are added to support JAVA™.
In view of the foregoing, it would be highly desirable to provide a technique for loading objects from a primary memory index that has a relatively small primary memory footprint. Such a technique could be used, for example, to reduce the primary memory footprint of the ZIP index for the core JAVA™ classes. Such a technique could also be used for applications that use large JAR files.
SUMMARY OF THE INVENTION
A method of loading objects from primary memory includes the step of creating hash values for listed objects to create a hash index specifying a set of hash values. The hash index is loaded into primary memory to form a primary memory hashed index object directory. A calculated hash value is derived for an invoked object. The calculated hash value is matched with a hash value in the primary memory hashed index object directory. In response to this matching operation, an object corresponding to the calculated hash value is loaded into primary memory.
A computer readable medium has an index generator with executable instructions to create hash values for listed objects, thereby creating a hash index specifying a set of hash values. The index generator loads the hash index into primary memory to form a primary memory hashed index object directory. An object loader has executable instructions to derive a calculated hash value for an invoked object. The object loader matches the calculated hash value with a hash value in the primary memory hashed index object directory. In response to a match, the object loader loads into primary memory an object corresponding to the calculated hash value.
The invention provides a technique for loading objects from a primary memory index that has a relatively small primary memory footprint. The technique of the invention is advantageously used to reduce the primary memory footprint of the ZIP index for the core JAVA™ classes. The technique can also be exploited for applications that use large JAR files.
REFERENCES:
patent: 5742807 (1998-04-01), Masinter
patent: 5754844 (1998-05-01), Fuller
patent: 5809494 (1998-09
Pennie & Edmonds LLP
Sun Microsystems Inc.
Verbrugge Kevin
LandOfFree
Apparatus and method for loading objects from a primary... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus and method for loading objects from a primary..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus and method for loading objects from a primary... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2577657