An appli­ca­tion net­work is a net­work of con­nected appli­ca­tions, data and devices, that sup­ports com­plex processes and work­flows. Appli­ca­tion net­works are often used, when com­pa­nies have very spe­cial­ized work­flows and can't find a sin­gle frame­work or soft­ware solu­tion to sup­port all its needs.

When an own mono­lithic appli­ca­tion might be too expen­sive to cre­ate, too hard to main­tain and extend, or just too unflex­i­ble to fastly react to mar­ket changes, the work­flows are bro­ken down into smaller steps and log­i­cal units. Each of these steps can then be sup­ported by smaller, spe­cial­ized and inter­linked appli­ca­tions or microser­vices. These appli­ca­tions may be cus­tom solu­tions or even stock soft­ware. To model the entire work­flows, they exchange its processed data with each other through stan­dard­ized APIs. The appli­ca­tion logic is deter­mined by the way how the data is processed and flows dur­ing the work­flows.

The approach of an appli­ca­tion net­work increases flex­i­bil­ity to the enter­prise. Once work­flows change, for exam­ple due to changed mar­ket con­di­tions, new appli­ca­tions and ser­vices can eas­ily be added to the net­work, old appli­ca­tions replaced, or the data flow and appli­ca­tion logic can be adapted.

As the appli­ca­tions and ser­vices that make up the appli­ca­tion net­work com­mu­ni­cate with each other using stan­dard­ized APIs, they can be devel­oped inde­pen­dently, be writ­ten in dif­fer­ent pro­gram­ming lan­guages, or hosted in ded­i­cated envi­ron­ments and cloud ser­vices. Rather than hav­ing a huge team work­ing on or main­tain­ing a sin­gle appli­ca­tion, devel­op­ment and main­te­nance can be split up into smaller teams, each respon­si­ble for sin­gle com­po­nents of the appli­ca­tion logic. Sin­gle fea­tures can be changed, extended and deployed faster.

And if the work­load of a ser­vice within the appli­ca­tion net­work is heav­ier than oth­ers, it can be scaled inde­pen­dently from the other ser­vices, allow­ing for a more effi­cient, to-the-point scal­ing of the entire appli­ca­tion net­work.