„The main aspect for some enterprises is to accelerate and streamline the business processes, while for others it is to restructure the organization, to build a more flexible architecture or secure the existing infrastructure.” – pointed out Zoltán Nagy, expert of Alerant Information Technologies Inc. 

Enterprise Service Buses
Restructuring the infrastructure according to the new approach is often inevitable and creates many challenges. The number of connections between services and systems are rocketed, at the same time the ability to centrally maintain, monitor and parameterize them is a basic requirement. Consequently an integration solution is needed that is also compliant with the SOA principles, thus ensuring the loose connection between service caller-and provider components, the reusability, interception and central management of services.
„The Enterprise Service Bus (ESB) is a component of the logical architecture that provides an integration solution in alignment with SOA principles.” – said Zoltán Nagy. The definitions of ESB do not specify further criteria, its actual role is determined by the individual usage instances. Therefore in case of SOA implementations the question of what type of ESB should be implemented is often raised. The most important expectation is that the solutions should support the standards, the message-or event-based communication, the message-and protocol transformations, the service-and message-level security as well as they should facilitate the content-based management and also the management and monitoring of SLA criteria. During the implementation of ESB – as an outcome of reusability – the enterprises are soon confronted with the problem that other applications also start using the services of the systems thus they must cope with a much more massive load as compared to initial estimations. Their protection therefore is essential.

Protection of the infrastructure
In a SOA architecture the caller systems rightfully expect that they are guaranteed to receive the response to their forwarded requests within the given time. Although the timeout management is a basic task of service buses, unique solutions are often required. It is typical for the caller systems that they specify in their forwarded message the time period for which they are „willing” to wait for the response and the background systems guarantee that they process the message within a time no longer than that. 
Another basic task of infrastructure protection solutions is to control the requests received from the caller applications thus protecting the background systems from being overloaded. The component that fulfils this task is often a separate instance within the architecture (gateway or gate keeper) but in practice it is also typical that this task is performed by the ESB or a developed solution implemented with the toolkit of ESB.
„Protection against overload should be practically implemented in a way that for each caller system and service we separately define the SLA criteria, such as time-based indicators or criteria related to throughput capacity.” – emphasized the expert. „Our experiences show that enterprises can enjoy the benefits derived from SOA architectures in case they precisely know their expectations and goals as early as at the time of implementation and they create a unique, customized solution in compliance with these. Initial decisions and selecting the right direction and strategy are of key importance.”