Your friend comes up with the following idea to optimize the performance of 2PC by reducing the time that transactions spend actively holding locks. In order to decide whether transaction T is ready to commit, the server checks if all the relevant locks are unlocked, but it doesn't actually lock them. If they are all unlocked, it writes Ilready TII, and responds to the coordinator with the Ilready Til message. Only when the coordinator responds with llcommit TII, does the server actually obtain the appropriate locks (using strict 2PL), execute the transaction, and finally flushes llcommit Til to the log. However, your company has an isolation requirement that conflicting transactions must always be committed in the same order across all servers in the cluster. E.g., T1 and T2 conflict, and if T1 is executed before T2 in server 1, T1 also has to be executed before T2 in server 2. Q5.1 5 Points Does the new proposal guarantee or break this isolation property? Explain your answer.