fragmentation
- InIP networks, fragmentation is the division of large data packets into several smaller data packets. Thus, when routing data packets over several networks, the packet length may exceed the maximum processable packet length of the routers or the permissible data packet length of the networks. This is the case, for example, when moving from Ethernet with a packet length of 1,518 bytes to X.25 with a packet length of 512 bytes. In such a case, the data packets are split into parts (fragmented) and sent individually over the network. Fragmentation is one of the standard functions of the IP protocol. Fragments are each provided with a complete IP header and transmitted as independent data packets. The fragments can take different paths over the network and reach the data station in different order. The data station must be able to sort the fragments and pass them in the correct order to the higher protocol layer. This is achieved by specifying an offset in the dedicated fragment offset field in the IP header. The process of assembling the individual data packets is called reassembling.
- In the case of storage media, fragmenting is the division of a data set during storage. In this process, the data is stored on the storage locations of a storage disk according to certain specifications of the storage organization. Fragmentation has a direct influence on the write speed and read speeds. If there is a lack of space, longer data sets can be stored on sectors that are distributed over the entire hard disk. This leads to a reduction in access time, which can be remedied by defragmentation.