Foundational MDA Patterns for Service-Oriented Computing .

Abstract

As the foundation of EDI, B2C and B2B, distributed computing is a key enabler for today's enterprises and will become even more important with the advent of cloud computing on the one hand and an ever more agile work organization on the other hand. Whilst the rapid evolution of distributed computing technologies in the last three decades has delivered a rich set of platforms and paradigms for building robust enterprise systems, it has also left a legacy of unresolved problems including fundamental inconsistencies between the concepts of the two leading distributed computing paradigms, i.e., distributed object computing and service-oriented architecture. Equally important, there is a growing need to lower the complexities and barriers involved in developing client applications, which range from large scale business applications and business processes to laptop programs and small apps on mobile devices. In this article, we present a unified conceptual framework for service-oriented computing based on the proven MDA (Model Driven Architecture) terminology stack. With the conceptual framework we consolidate, and give semantics to, core concepts of service-oriented computing and provide a set of foundational model transformation patterns that map between the presented concepts and further clarify practical service-oriented computing scenarios. Finally, we show how the developed framework perfectly fits to the OSM (Orthographical Software Modeling) approach.

Out-Of-Place debugging: a debugging architecture to reduce debugging interference

“…In this section we validate out-of-place debugging by evaluating IDRA with respect to performance and usability. In particular, we compare IDRA to TelePharo [20], the remote debugger for Pharo programs [31]. TelePharo can be seen as state-of-practice mainstream technology in remote debugging.…”

Mercury: Properties and Design of a Remote Debugging Solution using Reflection.

“…Bifrost is an extension to the Smalltalk MOP that relies on explicit meta-objects to provide sub-method [DDLM07] and partial behavioral reflection [TNCC03]. Bifrost is implemented through dynamic re-compilation of methods.…”

“…Bifrost is implemented through dynamic re-compilation of methods. Method invocations are intercepted using reflective methods [Mar06] similarly to the wrappers abstraction introduced by Brant et al [BFJR98]. These are subsequently recompiled using AST meta-objects that control the generated bytecode.…”

“…The benchmark is based on Tanter [TNCC03] and the Bifrost metrics are those reported in [Res12]. The benchmark measures the slowdown introduced by each solution for one million messages send to a test object when a) no instrumentation is present b) instrumentation is loaded but is disabled for this specific object and c) instrumentation is enabled on the test object of the micro-benchmark.…”

“…On line 1 we can see that the meta-action receives two arguments. The first argument (named reifications on our Script) represents meta-information relating to the event (such as the object that triggered a particular event), while the second argument reifies the reflectogram [TNCC03] of this particular meta-jump, which can control meta-level execution. On line 3 we instruct the execution on the remote target to halt in a context of the object that triggered the event.…”