Networking Reference
In-Depth Information
can be trusted and who should be distrusted. Sometimes there are central trusted
authorities, and sometimes trust computation is distributed to each agent. Sometimes
the agents emulate social and cognitive trust, but take care not to become vulnerable
because of it. Issues of concern in the design of these agents range from assessing
the intent of other nodes, whether they are selfish or not, how their propensity to
trust impacts their trust beliefs, how they incorporate evidence about trust into their
current beliefs [ 73 ], whether and how they forget about past evidence [ 51 , 52 , 78 ]and
so on. An important consideration is when the computed value can be considered a
measure of trust, but not a probabilistic estimate of reliability [ 23 ]. Another concern
is making sure that computed trust value is useful for the agent's function, which is
not necessarily a conscious concern in the cognitive notion of trust.
The term trust is routinely used in computer networking to refer to this type of
measures computed by agents. An agent situated at a computational node observes
the success and failure of packets routed through other nodes. It sends recommen-
dations about other nodes in the network to help them form trust impressions about
their neighbors [ 14 , 30 ]. These methods incorporate algorithms to establish how
much the trustor should trust the trustee (trust computation and update), which trust
information should be sent to other nodes (recommendations) and how to inform
other nodes when a node should no longer be trusted (trust revocation). These
actions taken together constitute a trust management scheme. Trust computation
enhances the existing routing methods by routing through trusted peers.
Trust in Information
Trust in information has been studied extensively on the Internet. The exchange
of information is one of the primary uses of the Internet. Early research in this area
investigated the signals users rely on to judge the credibility of content found online.
We discussed research in this area in detail in Sect. 3.3 . In particular, the look and
feel of a site serves an important role in the formation of quick judgments about
whether the site is professional and contains current information. Various seals of
approval provided by outside institutions, help enforce this view [ 53 ]. The user's
expertise in the search area and the sites' relevance to the topic are among the
many factors that determine a user's trust based on the site's content [ 27 ]. The
reputation of the sources also contribute to the judgments for trust in content. A
site's reputation can be based on its ranking in search engines, links from other
sites, or other recommendations. However, this information is replaced by the user's
own judgment as she gains experience with the site [ 25 ](Fig. 4.2 ).
Analysis of a site's content is costly in terms of cognitive resources, and is
likely to replaced by heuristics based on secondary cues like appearance [ 36 ]. In
some instances, the individual lacks either the desire or the ability to engage in
systematic processing of content. However, in other cases, the individual is engaged
in a purposeful effort to understand a specific topic by building a mental model of
what information is available [ 63 ]. A quick review of available sites is used to judge
Search MirCeyron ::

Custom Search