The SALSA Recommendation IP5, Essay Example

SALSA is abbreviated as Simple Actor Language, System, and Architecture and is considered as an actor programming language consisting of high level support for remote messaging, universal naming, migration and coordination (Wyrzykowski, Dongarra, & Paprzycki, n.d). Likewise, the actor provides encapsulation for both states i.e. procedures and data along thread and control. The communication to and from actors is via asynchronous message transmissions. However, in between messaging process, an actor is capable of performing three types of basic operations i.e. modifying its own state, establishing new actors and sending messages to subordinate actors. Thus, in SALSA, actors are self-directed, synchronized and most importantly independent resulting in an effective process for parallel execution along with the support of mobility. SALSA is programmed on Java, supplementing the presence of deploying virtual machines on multiple heterogeneous software and hardware architectures (Wyrzykowski, Dongarra, & Paprzycki, n.d).

Moreover, the SALSA framework comprises of a set of classes for augmenting the deployment of an agent’s internal architecture along with taking control of the life cycle. Likewise, the SALSA application programming interface was first deployed in Java programming language that facilitates agents to execute regardless of any platform. Although, some of SALSA versions do not activate or allow agents to contact native libraries associated with Windows, as a result, a subset of SALSA APL was introduced programmed in C# that is called Micro SALSA (mSALSA) (Sordo, Yoshida, Vaidya, & Jain, n.d). mSALSA facilitate programmers to construct agent components and utilize the communication protocol of SALSA. Moreover, programmers also have the facility of coding in any of the programming languages that will beneficial (Sordo, Yoshida, Vaidya, & Jain, n.d).

There are two types of observations that were recognized for SALSA agents i.e. active and passive. The active observation requires permission for making a decision of requesting or gathering information from any other agent. However, in passive observation, data is transferred to the agent without a request (Sordo, Yoshida, Vaidya, & Jain, n.d). SALSA supports reconfigurable supported strategies at different levels apart from some constraints that are demonstrated below (Chang, 2007):

SALSA is implemented simultaneously with Java class libraries. Likewise, the objects that be relocated adds more features via special design class (Chang, 2007). However, overhead for small components is high and communication held locally is not exposed.

The SALSA distributed strategy is only applicable to objects, as preliminary nodes can be allocated and strategies for migration can be quantified as well (Chang, 2007). However, there is not much control on loading the code as the system can assume a shared code or an installation on all available nodes (Chang, 2007). Likewise, reconfigurable distribution of the code can only be accomplished via customized Java class loaders.

References

Wyrzykowski, R., Dongarra, J., & Paprzycki, M. Parallel processing and applied mathematics: 5th international conference, PPAM 2003, czestochowa, poland, september 7-10, 2003. revised papers (lecture notes in computer science) Springer.

Sordo, M., Yoshida, H., Vaidya, S., & Jain, L. C. Advanced computational intelligence paradigms in healthcare Berlin ; Springer, 2007-.

Chang, P. (2007). Transparent distributed computing. _{Customizing web applications through adaptable components and reconfigurable distribution}