Update final-exam-review-questions.md

operating-systems/final-exam-review-questions.md

@@ -123,7 +123,7 @@ Basically, we need atomic operations because the checking of a lock and the sett
 
 ### Why are spinlocks useful? Would you use a spinlock in every place where you're currently using a mutex?
 
-On one hand, spinlocks are very simple to implement. There is use in simplicity. In addition, they are pretty simple to understand. You just burn through CPU cycles continuously checking the value of the lock until you are preempted. As well, spinlocks have very low latency and very low delay. You probably wouldn't want to use spin locks every place you are currently using a mutex, however. What's nice about mutexes is that, after checking a lock that isn't free, the process will immediately relinquish the CPU. This can have some performance considerations. That being said, when a spin lock is relinquished there is pretty much no delay in that link being acquired by the next in line.
+On one hand, spinlocks are very simple to implement. There is use in simplicity. In addition, they are pretty simple to understand. You just burn through CPU cycles continuously checking the value of the lock until you are preempted. As well, spinlocks have very low latency and very low delay. You probably wouldn't want to use spinlocks every place you are currently using a mutex, however. What's nice about mutexes is that, after checking a lock that isn't free, the process will immediately relinquish the CPU. This can have some performance considerations. That being said, when a spinlock is relinquished there is pretty much no delay in that link being acquired by the next in line.
 
 ### Do you understand why is it useful to have more powerful synchronization constructs, like reader-writer locks or monitors? What about them makes them more powerful than using spinlocks, or mutexes and condition variables?
 
@@ -221,7 +221,7 @@ An RPC system will have to specially encode variable length arrays. In SunRPC, v
 
 There are many different design options when implementing a distributed service.
 
-On one hand, everything can be driven through the client. This is nice because it makes the server super simple. Unfortunately, the client now has to download and re-upload all files, even for very small modifications. In addition, the server completely loses control over files. On the other, hand everything can be driven through the server. This is nice because the server retains control over the files. Unfortunately, every file operation must incur a network cost, which limits the scalability of the server.
+On one hand, everything can be driven through the client. This is nice because it makes the server super simple. Unfortunately, the client now has to download and re-upload all files, even for very small modifications. In addition, the server completely loses control over files. On the other hand, everything can be driven through the server. This is nice because the server retains control over the files. Unfortunately, every file operation must incur a network cost, which limits the scalability of the server.
 
 Replication can be a powerful technique. Replication involves duplicating files across all nodes in the system. This is helpful because any node can service any file request. If a node goes down, any other node can be used. This makes replicated systems highly available and fault tolerant. Unfortunately, these systems are harder to scale, since every node needs to hold all of the data.
 
@@ -294,7 +294,7 @@ In the weak consistency model, the synchronization effort is placed back on the
 
 In a homogenous design, all of the nodes in the system are able to fulfill any request that enters the system. The benefit of this design is that the frontend load-balancing component can be simple. It simply needs to send requests to each of the nodes in a round-robin manner. Unfortunately, this design doesn't allow for any benefit from caching. The frontend component doesn't keep enough state in this design to understand the locality of tasks on a node-by-node basis.
 
-In a heterogenous design, each node is specialized to fulfill some subset of requests that enter the system. The benefit of this design is that caching can be leveraged more effectively. With similar computations being performed on similar data, the possibility of actionable data being present in the cache increases. Our system has more locality. The frontend becomes more complex, as it now needs to know more about the system in order to route requests. A big downside is that the scaling opportunities become more complex. It's not longer enough to add or remove nodes in response to fluctuating traffic. In heterogenous systems, you'll need to understand which components of the system need to be scaled up or scaled down.
+In a heterogenous design, each node is specialized to fulfill some subset of requests that enter the system. The benefit of this design is that caching can be leveraged more effectively. With similar computations being performed on similar data, the possibility of actionable data being present in the cache increases. Our system has more locality. The frontend becomes more complex, as it now needs to know more about the system in order to route requests. A big downside is that the scaling opportunities become more complex. It's no longer enough to add or remove nodes in response to fluctuating traffic. In heterogenous systems, you'll need to understand which components of the system need to be scaled up or scaled down.
 
 ### Do you understand the history and motivation behind cloud computing, and basic models of cloud offerings? Do you understand some of the enabling technologies that make cloud offerings broadly useful?