Page 110 - Handbook of Deep Learning in Biomedical Engineering Techniques and Applications
P. 110

98   Chapter 4 A critical review on using blockchain technology in education domain




                                    3.3 Directed acyclic graph approaches
                                    In a BC, blocks can be added sequentially to a single linear chain,
                                    after a waiting time. Directed acyclic graph (DAG) technology adds
                                    single transaction (not blocks) to multiple branches of a tree struc-
                                    ture in parallel. A DAG is a tree that links a transaction to multiple
                                    transactions. DAG is a public permissioned noneBC-based distrib-
                                    uted ledger technology. IOTA and Byteball use this approach. The
                                    main advantage of DAGs is its speed and scalability. The possibility
                                    of double spending in BCs is eliminated as transactions are
                                    validated based on the previous transactions linked to it.
                                    Challenges include centralization due to preselection of validators
                                    and forming of parasite chains on invalid transactions. BC
                                    consensus requires validation of blocks (with multiple transac-
                                    tions). In DAG, individual nodes act as both miners and validators,
                                    as one transaction validates another [14]. Table 4.3 compares the
                                    features of linear BC with tree-based DAG for digital ledgers.
                                       Some of the advantages of DAG-based approach include the
                                    following:
                                    • Overcoming double spending as DAG validates a particular
                                       transaction based on the previous number of transactions
                                    • Reducing transaction size by linking DAG to existing graphs



                   Table 4.3 Comparison of BC-based DLT and DAG-based DLT approaches.

                BC-based DLT                     DAG-based DLT

                Linear chain of blocks (set of transactions)  Tree of transactions
                Validation of blocks             Validation of transactions
                Public/private/consortium        Public permissioned
                Medium volume of transactions    Large volumes of transactions (micro/nanotransactions)
                Transparency                     Less transparent
                Immutable                        Immutable
                High latency                     Low latency
                High energy consumption (PoW)    Low energy consumption
                Higher user fees                 Avoids user fees
                Scalable                         Highly scalable
                Tends to become centralized (power  More decentralized
                 concentration)
                Not so                           Susceptible to security lapse if volume of transactions is
                                                   reduced
                BC, blockchain; DAG, directed acyclic graph; DLT, digital ledger technology; PoW, proof of work.
   105   106   107   108   109   110   111   112   113   114   115