Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
2000-06-07
2003-05-27
Popovici, Dov (Department: 2175)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000, C707S793000, C709S202000, C711S124000, C711S146000
Reexamination Certificate
active
06571257
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to the management of computer storage infrastructures, and more particularly to the management of storage devices as resources i.e., Storage Resource Management (SRM).
BACKGROUND OF THE INVENTION
Data storage is becoming increasingly important to Information Technology (IT) organizations. Spending on storage, such as disks, tapes, other storage related hardware and software, typically consumes 50% of IT spending. This situation has created a need for software to manage this storage infrastructure particularly in a network environment.
There are two broad categories of software used for storage management: (1) Data Management and (2) Storage Resource Management (SRM). Data Management software focuses on protecting the data stored. This software includes traditional backup applications, as well as other mirroring and replication solutions. SRM software focuses on managing storage devices as resources, independent of the data itself. This software includes capabilities for capacity planning and allocation, logical volume management, and storage related performance monitoring. The present patent document focuses on Storage Resource Management.
One key problem which all SRM applications must contend with is obtaining accurate, timely information about the storage resources attached to a computer. Applications must have information regarding, for example, file ownership by user and application, aggregate usage information for files by user and application, allocation of storage on devices such as high speed disks, mirrored disks, tape, etc., performance, and throughput information, etc. In addition, modem computer systems often include networks of computers which range from only a few to over 10,000 machines. SRM applications need to be able to easily scale between networks of these sizes.
Currently there are two ways in which SRM applications collect information, such as mention above, from a computer. These methods are (1) batch data collection and (2) Operating System (OS) obtained information.
With batch data collection, a process on a computer periodically examines every file on the computer's file system, recording SRM information on a file-by-file basis. The primary advantage of this technique is that it can provide complete information on all the files on a computer. This information can be gathered together to provide aggregate information by user and/or application. But, this technique has some serious disadvantages. First the technique is very slow. Each file on the system must be examined. As a result, the more files a system has, the more time it takes to scan all the files. The fact that it must examine every file means that there will be a significant impact on the performance of the system when the application runs. These performance issues limit the use of this technique to less than a few hundred computer systems which is far short of the true scalable needs. A further limitation of this technique is that it cannot collect performance related information. The data that it collects merely provides a snapshot of the system at specified times. An example of a product which uses batch data collection is “SAME: Vantage” from Sterling Software.
The other technique used to collect SRM related information is to take advantage of information which can be obtained from the Operating System, such as configuration information and data from performance counters. While this technique is very fast and scalable, the types of information it collects are limited. For example, it cannot provide any information on a file-by-file basis. Examples of products that use Operating System information are system tools “Top & Monitor” on HP-UX and the Windows NT “Performance Monitor”, as well as High ground's “Storage Resource Manager”.
Thus, there is a need for an SRM system which can provide the detailed information which batch data collection techniques do without the associated performance penalties which severely limit the ability to scale of the network in which storage is being managed, while at the same time providing the performance information obtainable from the Operating System.
SUMMARY OF THE INVENTION
The present patent document discloses methods for Storage Resource Management which can be easily and efficiently scaled to computer systems comprising many thousands of computers while providing only detailed file storage attributes. The method disclosed in the present patent document comprises three phases: (1) an initialization phase, (2) a data collection phase, and (3) a management server communication phase.
During the initialization phase, descriptive data, referred to herein as file storage attributes, is collected in order to create a snapshot of the storage layout on every managed host, which could be for example by means of a batch data collection method. The managed hosts are computers, typically attached to a network, whose storage is to be managed.
During the collection phase calls between user applications and the file system are intercepted. Key file data attributes are recorded in a buffer and subsequently read and transferred to a local data repository. The information transferred represents changes to the file system and is merged with the baseline data-set collected during the initialization phase. This merge step ensures that the contents of the local data repository reflect the true state of the file system. The incremental nature of the updates provide a key advantage of the invention - it can provide a real-time view of the file system without having to resort to batch data collection methods.
During the communication phase information is transferred from the local data repository to a storage management application located on a storage management server. The storage management application is responsible for aggregating storage information across multiple managed hosts and presenting results to the system manager.
A primary advantage of the embodiment as described in the present patent document over prior solutions lies in the fact that the system is scalable to systems comprising computers in excess of a few hundred. It collects complete information because it intercepts all calls relating to file operations but has minimal impact on system performance. Following the initialization phase, it does not have to scan each file because it is inserted into the normal flow of file operations from applications to the file system. The system has immediate information regarding any changes to files. Further, it can collect performance information on a per file basis. It is fast and scalable to many systems.
REFERENCES:
patent: 5495607 (1996-02-01), Pisello et al.
patent: 5666532 (1997-09-01), Saks et al.
Duggan Gerald P
Poppen-Chambers Ronald E
Popovici Dov
Solaiman Shireen I
LandOfFree
Efficient, real-time management of storage resources does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Efficient, real-time management of storage resources, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Efficient, real-time management of storage resources will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3074910