tured paradigm, well anchored in the present Internet.
On the other hand, we have pointed out that the Fu-
ture Internet, perceived as the evolution of the current
Internet, has to face existing challenges (i.e., related to
scalability, heterogeneity, mobility, awareness & adapt-
ability, and security, privacy & trust), while pushed to
the extreme, as summarized in Table 2. Indeed, the
Internet of Services and Things is becoming a reality
with a population of services, including service-enabled
things, bound to evolve at a very fast pace. Then, as-
sisting the developers in leveraging such a plethora of
services to provide new applications raises the need to
carefully revisit the service-oriented middleware solu-
tions developed for today's Internet (see Table 4).
Overall, the middleware must cope with the trade-
o between the increasing scale and heterogeneity of
the Future Internet re
ected on the information ex-
posed by service descriptions and the complexity of
processing this information, to enable service publica-
tion, discovery, composition and access. With respect
to service publication and discovery, the correct ar-
chitectural choices must be made for the correspond-
ing publication and discovery protocols, to balance the
trade-o introduced among keeping the complexity of
the protocols low, handling the dierent dimensions of
the Future Internet scale/heterogeneity/mobility, con-
trolling the quality of the information managed by
the protocols, and providing reasonable security/pri-
vacy/trust guarantees. From the standpoint of service
access, the middleware must benet as much as pos-
sible from high-performance computing paradigms and