386 SYNCHRONOUS COMPUTATIONS We are going to consider all three…

Question Answered step-by-step 386 SYNCHRONOUS COMPUTATIONS We are going to consider all three… 386SYNCHRONOUS COMPUTATIONSWe are going to consider all three problems and examine their nature and interplayin some details. All of them will be considered under the standard set of restrictionR plus obviously Synch.6.4.1 Reset/Wake-UpIn reset, all entities must enter the same state within finite time. One important application of reset is when a distributed protocol is only initiated by a subset of the entitiesin the system, and we need all entities in the system to eventually begin executingthe protocol. When reset is applied at the first step of a protocol, it is called wake-up.The wake-up or reset problem is a fundamental problem and we have extensivelyexamined in asynchronous systems.In fully synchronous systems it is sometimes also called weak unison; its solutionis usually a preliminary step in larger computations (e.g, Wait, Guess, DoubleWait),and it is mostly used to keep the initiation times of the main computation bounded. Forexample, in protocol Wait applied to a network G (not necessarily a ring) of knowndiameter d, the initial wake-up ensures that all entities become awake within d timeunits from the start.For computations that use wake-up as a tool, their cost obviously depends on thecost of the wake-up. Consider for example electing a leader in a complete graph Knusing the waiting technique. Not counting the wake-up, the election will cost onlyn ? 1 bits, and it can be done in 4imin + 1 time units (see Equation 6.29); recall thatin a complete graph,d = 1. Also, the wake-up can be done fast, in 1 time unit, butthis can cost O(n2) bits. In other words, the dominant bit cost in the entire electionprotocol is the one of the wake-up, and it is unbearably high. Sometimes it is desirableto obtain wake-up protocols that are slower but use fewer transmissions.In the rest of this section we will concentrate on the problem of wake-up in acomplete network. The difficulty of waking up in asynchronous complete networks,which we discussed in Section 2.2, does not disappear in synchronous completenetworks. In fact, in complete networks where the port numbers are arbitrary, (n2)signals must be sent in the worst case.Theorem 6.4.1 In a synchronous complete network with arbitrary labeling, wake-uprequires (n2) messages in the worst case.To see why this is true, consider any wake-up protocol W that works for anycomplete networks regardless of the labeling. By contradiction, letW use o(n2) signalsin every complete network of size n.We will first consider a complete network Kn1 with chordal labeling: A Hamiltoniancycle is identified, and a link (x, y) is labeled with the distance from x to y accordingto that cycle. The links incident on x will, thus, be labeled Image transcription textJ). For the following LP model, anywer the questions. Min. RX – 12YIX – JY29 2X + 2Y 2 10 and X. Y20 a). Use the graphical solutionprocedure to find optimal solution. bi Assume that the… Show more… Show moreImage transcription textPROBLEMS, SECTION 14 Evaluateeach of the following in r + ly form, andcompare with a computer… Show more… Show more  Computer Science Engineering & Technology Java Programming COMPUTER computer Share QuestionEmailCopy link Comments (0)