Chapter 2 checkpoint

Chapter 2 checkpoint: availability, consistency and estimation

Every question from this chapter’s lessons. Your first answer to each is the one that counts; afterward you can practice again as often as you like. Passing needs 75% right on the first attempt.

0 of 16 answered0 right on the first attempt

Question 1 of 16

Which consistency pattern guarantees that, once a write completes, every later read sees it?

Eventual consistencyStrong consistencyMonotonic readsWeak consistency

Question 2 of 16

Match each consistency pattern to the kind of system the lesson says uses it.

Strong consistencyEventual consistencyWeak consistency

Question 3 of 16

A user updates their profile, which is written to the primary database. The page then reloads from a read replica that copies the primary asynchronously, and shows the old profile. What is happening?

The replica has not received the write yet, so the read is staleOnly a browser cache could cause thisThe write was lostThe two databases are deadlocked

Question 4 of 16

A Dynamo-style store keeps

N = 5

copies of each item and waits for

W = 3

acknowledgments on every write. What is the smallest read quorum

R

that makes every read overlap every write?

Your answerreplicas

Answers within 1% count.

Question 5 of 16

A database runs as two servers in an active-passive pair. Why can it handle only as much traffic as one server?

Each server runs at half speed to stay in stepOnly the active server handles traffic; the passive one waits to take overBoth servers must write to the same diskThe heartbeats use up half of each server’s capacity

Question 6 of 16

An active-active web tier must carry a peak of 9,000 requests per second, and each server handles 2,000. How many servers does it need so that losing any one still covers the peak?

Your answerservers

Answers within 1% count.

Question 7 of 16

The active server of an active-passive pair crashes, and the passive one takes over within seconds. Afterward, a few orders placed just before the crash are missing. What is the most likely cause?

Those writes had not been replicated to the passive server when the active one failedThe passive server was a cold standbyThe heartbeats were sent too oftenDNS sent the orders to the wrong server

Question 8 of 16

What does master-master replication offer that master-slave does not, and what does it cost?

It removes write conflicts altogetherWrites continue if one master fails; you need routing for writes and risk conflicts or slower writesIt lets the system run read-only while a replica is promotedIt limits capacity to a single server

Question 9 of 16

A request passes through three tiers in series: DNS at 99.99%, a load balancer at 99.9% and an application server at 99.9%. What is the availability of the whole path, as a percentage?

Your answer%

Answers within 0.05% count.

Question 10 of 16

Two independent servers each have 99% availability, and either one can serve all requests. What is the availability of the pair, as a percentage?

Your answer%

Answers within 0.005% count.

Question 11 of 16

A service promises 99.9% availability. How many minutes of downtime does that allow in an average month?

Your answer

Answers within 5% count, in any unit you choose.

Question 12 of 16

Two redundant servers, each 99.9% available, should together give 99.9999%. But both went down at once when the power supply they share failed. What did the calculation get wrong?

Redundant servers should be multiplied as a seriesTwo servers are never enough for six ninesAvailability should have been measured per requestIt assumed the two servers fail independently

Question 13 of 16

An application receives 86.4 million requests a day, and at its peak traffic is 2.5 times the daily average. What is the peak in requests per second?

Your answerRPS

Answers within 2% count.

Question 14 of 16

One million users each post five photos a day, and a photo is 500 KB. Without compression, how many terabytes do three years of photos need?

Your answer

Answers within 10% count, in any unit you choose.

Question 15 of 16

Put these operations in order from fastest to slowest, using the latency numbers in the lesson.

L1 cache reference
Round trip within a datacenter
Main memory reference
Hard disk seek
Packet California → Netherlands → California
Read 4 KB randomly from SSD

Question 16 of 16

About how many requests per second is one billion requests a day, on average?

Your answerRPS

Answers within 5% count.