Sequential access is a process used for retrieving data from a storage device. It is also known as serial access. In sequential access, the storage device moves through all information up to the point it is attempting to read or write. An example of sequential access drive is a tape drive where the drive moves the tape forward or backward until the destination is reached. Sequential access memory can also be called "storage system." The data is stored and read in a sequential fixed order. Sequential access is the type of memory mostly used for permanent storage, whereas, random access memory is used for temporary storage.
A data structure can be accessed sequentially if one can visit the values in one particular order. The best example is the linked list. Indexing into a list having sequential access takes O (k) time, where k is the index. As a result, many algorithms such as quick sort and binary search degenerate into bad algorithms. They are less efficient than their naive alternatives. These algorithms are not practically possible without random access. On the other hand, index such as merge sort algorithm requires only sequential access. It works best for text files created with a typical text editor. Sequential access is not the right choice for long series of numbers stored as a character string. Hence, a four-digit number would require four bytes of storage whereas; if the number is stored as an integer, it would require only two bytes of storage.
Old recording media such as CDs, DVDs, and magnetic tapes are examples of sequential access memory drives. Hard drive is also an example of sequential access memory. Examples of random access memory include memory chips and flash memory (such as memory sticks or memory cards).
Comparing sequential versus random disk operations helps to assess systems efficiency. Accessing data sequentially is faster than random operations, because it involves more search functions. The search operation is performed by the right disk cylinder. It occurs when the disk head positions itself to access the data requested for. More ever, random access delivers a lower rate of output. If the disk access is random, it is advisable to pay attention and monitor for the emergence of any bottleneck. For workloads of either random or sequential input/output, it is advisable to use drives with faster rotational speeds. For workloads that are predominantly random input/output, it is advisable to use a drive with faster search time.
The number of records that are affected when updating a file refers to its hit rate. Let us consider a file with 5000 records; if there is a delete or an update operation affecting only 50 records, then the hit rate is very low. If there are 4500 records that are affected by update or delete operations, then the hit rate is high. Sequential access is found to be slow when the hit rate is low. It is due to the fact that sequential access has to search all the records in a particular order. Moreover, sequential files are executed in a batched transaction to overcome the problem of low hit rate. This makes the sequential access more complicated than direct file access.