You are on page 1of 4

Direct Access File System (DAFS

Direct Access File System (DAFS) is a file-access sharing protocol that uses
memory-to-memory interconnect architectures, such as VI and InfiniBand. DAFS is
designed for storage area networks (SANs) to provide bulk data transfer directly
between the application buffers of two machines without having to packetize the
data. With DAFS an application can transfer data to and from application buffers
without using the operating system, which frees up the processor and operating
system for other processes and allows files to be accessed by servers using several
different operating systems

random access
Refers to the ability to access data at random. The opposite of random access is sequential
access. To go from point A to point Z in a sequential-access system, you must pass through all
intervening points. In a random-access system, you can jump directly to point Z. Disks are
random access media, whereas tapes are sequential access media.

The terms random access and sequential access are often used to describe data files. A random-
access data file enables you to read or write information anywhere in the file. In a sequential-
access file, you can only read and write information sequentially, starting from the beginning of
the file.

Both types of files have advantages and disadvantages. If you are always accessing information
in the same order, a sequential-access file is faster. If you tend to access information randomly,
random access is better.

Random access is sometimes called direct access.

The Direct Access File System (DAFS) is a new, fast, and lightweight remote file system
protocol. DAFS targets the data center by addressing the performance and functional needs of
clusters of application servers. We call this the local file sharing environment. File access
performance is improved by utilizing Direct Access Transports, such as InfiniBand, Remote
Direct Data Placement, and the Virtual Interface Architecture. DAFS also enhances file sharing
semantics compared to prior network file system protocols. Applications using DAFS through a
user-space I/O library can bypass operating system overhead, further improving performance.
We present performance measurements of an IP-based DAFS network, demonstrating the DAFS
protocol's lower client CPU requirements over commodity Gigabit Ethernet. We also provide the
first multiprocessor scaling results for a well-known application (GNU gzip) converted to use
Direct or random-access data file

A random-access data file enables you to read or write information anywhere in the file

sequential-access file

In a sequential-access file, you can only read and write information sequentially, starting from
the beginning of the file.

What is file organization? What are the

various file organizational methods?
File organization refers to the relationship of the key of the record to the physical location
of that record in the computer file.
File organization may be either physical file or a logical file. A physical file is a physical unit,
such as magnetic tape or a disk.
A logical file on the other hand is a complete set of records for a specific application or purpose.
A logical file may occupy a part of physical file or may extend over more than one physical file.

The objectives of computer based file organization:

 Ease of file creation and maintenance

 Efficient means of storing and retrieving information.

The various file organization methods are:

 Sequential access.
 Direct or random access.
 Index sequential access.

The selection of a particular method depends on:

 Type of application.
 Method of processing.
 Size of the file.
 File inquiry capabilities.
 File volatility.
 The response time.
1. Sequential access method: Here the records are arranged in the ascending or descending
order or chronological order of a key field which may be numeric or both. Since the records are
ordered by a key field, there is no storage location identification. It is used in applications like
payroll management where the file is to be processed in entirety, i.e. each record is processed.
Here, to have an access to a particular record, each record must be examined until we get the
desired record.
Sequential files are normally created and stored on magnetic tape using batch processing method.


 Simple to understand.
 Easy to maintain and organize
 Loading a record requires only the record key.
 Relatively inexpensive I/O media and devices can be used.
 Easy to reconstruct the files.
 The proportion of file records to be processed is high.


 Entire file must be processed, to get specific information.

 Very low activity rate stored.
 Transactions must be stored and placed in sequence prior to processing.
 Data redundancy is high, as same data can be stored at different places with different keys.
 Impossible to handle random enquiries.

2. Direct access files organization: (Random or relative organization). Files in his type are
stored in direct access storage devices such as magnetic disk, using an identifying key. The
identifying key relates to tits actual storage position in the file. The computer can directly locate
the key to find the desired record without having to search through any other record first. Here
the records are stored randomly, hence the name random file. It uses online system where the
response and updation are fast.


 Records can be immediately accessed for updation.

 Several files can be simultaneously updated during transaction processing.
 Transaction need not be sorted.
 Existing records can be amended or modified.
 Very easy to handle random enquiries.
 Most suitable for interactive online applications.


 Data may be accidentally erased or over written unless special precautions are taken.
 Risk of loss of accuracy and breach of security. Special backup and reconstruction
procedures must be established.
 Les efficient use of storage space.
 Expensive hardware and software are required.
 High complexity in programming.
 File updation is more difficult when compared to that of sequential method.

3. Indexed sequential access organization: Here the records are stored sequentially on a direct
access device i.e. magnetic disk and the data is accessible randomly and sequentially. It covers
the positive aspects of both sequential and direct access files.
The type of file organization is suitable for both batch processing and online processing.
Here, the records are organized in sequence foe efficient processing of large batch jobs but an
index is also used to speed up access to the records.
Indexing permit access to selected records without searching the entire file.


 Permits efficient and economic use of sequential processing technique when the activity rate
is high.
 Permits quick access to records, in a relatively efficient way when this activity is a fraction of
the work load.


 Slow retrieval, when compared to other methods.

 Does not use the storage space efficiently.
 Hardware and software used are relatively expensive

Choosing a file organization is a design decision, hence it must be done having in mind the
achievement of good performance with respect to the most likely usage of the file. The criteria
usually considered important are:

1. Fast access to single record or collection of related recors.

2. Easy record adding/update/removal, without disrupting (1).
3. Storage efficiency.
4. Redundance as a warranty against data corruption.