David Clinton - AWS Certified Solutions Architect Study Guide

9 B. S3 uses a read‐after‐write consistency model for new objects, so once you upload an object to S3, it's immediately available. See Chapter 3for more information.

10 C. You can't change the primary CIDR for a VPC, so you must create a new one to connect it to your internal network. See Chapter 4for more information.

11 B. An EC2 instance can access the Internet from a private subnet provided it uses a NAT gateway or NAT instance. See Chapter 4for more information.

12 A. The definition of a public subnet is a subnet that has a default route pointing to an Internet gateway as a target. Otherwise, it's a private subnet. See Chapter 4for more information.

13 C. DynamoDB is a key‐value store that can be used to store items up to 400 KB in size. See Chapter 5for more information.

14 A.You can create a global secondary index for an existing table at any time. You can create a local secondary index only when you create the table. See Chapter 5for more information.

15 A. Enabling point‐in‐time recovery gives you an RPO of about five minutes. The recovery time objective (RTO) depends on the amount of data to restore. See Chapter 5for more information.

16 B. Revoking unnecessary access for IAM users is the most effective of the listed measures for protecting your AWS account. See Chapter 6for more information.

17 C. KMS can be used to encrypt Elastic Block Store (EBS) volumes that store an instance's operating system. See Chapter 6for more information.

18 D. STS tokens expire and IAM access keys do not. An STS token can be used more than once. IAM access keys and STS tokens are both unique. An IAM principal can use an STS token. See Chapter 6for more information.

19 B. EC2 doesn't track instance memory utilization. See Chapter 7for more information.

20 C. The transition to the ALARM state simply implies that the metric crossed a threshold but doesn't tell you what the threshold is. Newly created alarms start out in the INSUFFICIENT_DATA state. See Chapter 7for more information.

21 A. Both store their logs in S3 buckets. See Chapter 7for more information.

22 A. An EC2 instance in a private subnet still has access to Amazon's private DNS servers, which can resolve records stored in public hosted zones. See Chapter 8for more information.

23 C. Geoproximity routing routes users to the location closest to them. Geolocation routing requires you to create records for specific locations or create a default record. See Chapter 8for more information.

24 A. Route 53 is a true DNS service in that it can host zones for any domain name. You can also register domain names with or transfer them to Route 53. See Chapter 8for more information.

25 B. Lambda is a highly available, reliable, “serverless” compute platform that runs functions as needed and scales elastically to meet demand. EC2 spot instances can be shut down on short notice. See Chapter 10for more information.

26 A. A simple scaling policy changes the group size and then has a cooldown period before doing so again. Step scaling policies don't have cooldown periods. Target tracking policies attempt to keep a metric at a set value. PercentChangeInCapacity is a simple scaling adjustment type, not a scaling policy. See Chapter 10for more information.

27 A. Auto Scaling always attempts to maintain the minimum group size or, if set, the desired capacity. See Chapter 10for more information.

28 D. ElastiCache supports Memcached and Redis, but only the latter can store data persistently. See Chapter 11for more information.

29 B. Puppet is a configuration management platform that AWS offers via OpsWorks but is not itself an AWS service. See Chapter 11for more information.

30 B. S3 cross‐region replication transfers objects between different buckets. Transfer acceleration uses a CloudFront edge location to speed up transfers between S3 and the Internet. See Chapter 11for more information.

31 A. You can deactivate STS for all regions except US East. See Chapter 12for more information.

32 A. GuardDuty looks for potentially malicious activity. Inspector looks for vulnerabilities that may result in compromise. Shield and Web Application Firewall protect applications from attack. See Chapter 12for more information.

33 A. Applying encryption to an unencrypted object will create a new, encrypted version of that object. Previous versions remain unencrypted. See Chapter 12for more information.

34 C. On‐demand instances will continue to run and incur costs. Reserved instances cost the same whether they're running or stopped. Spot instances will be terminated when the spot price exceeds your bid price. See Chapter 13for more information.

35 A. The EBS Lifecycle Manager can take scheduled snapshots of any EBS volume, regardless of attachment state. See Chapter 13for more information.

36 C. Elastic Container Service lets you run containers that can launch in a matter of seconds. EC2 instances take longer. Lambda is “serverless,” so you can't use it to run a web server. CloudFront provides caching but isn't a web server. See Chapter 13for more information.

37 A. Almost everything in CloudFormation is case sensitive. See Chapter 14for more information.

38 A, C. CodeDeploy looks for the appspec.yml file with the application files it is to deploy, which can be stored in S3 or on GitHub. See Chapter 14for more information.

39 B. You can use CodeDeploy to deploy an application to Lambda or EC2 instances. But an AWS Systems Manager command document works only on EC2 instances. See Chapter 14for more information.

The cloud is where much of the serious technology innovation and growth happens these days, and Amazon Web Services (AWS), more than any other, is the platform of choice for business and institutional workloads. If you want to be successful as an AWS solutions architect, you'll first need to understand what the cloud really is and how Amazon's end of it works.

TO MAKE SURE YOU'VE GOT THE BIG PICTURE, THIS CHAPTER WILL EXPLORE THE BASICS:

What makes cloud computing different from other applications and client‐server models

How the AWS platform provides secure and flexible virtual networked environments for your resources

How AWS provides such a high level of service reliability

How to access and manage your AWS‐based resources

Where you can go for documentation and help with your AWS deployments

The technology that lies at the core of all cloud operations is virtualization. As illustrated in Figure 1.1, virtualization lets you divide the hardware resources of a single physical server into smaller units. That physical server could therefore host multiple virtual machines (VMs) running their own complete operating systems, each with its own memory, storage, and network access.

FIGURE 1.1 A virtual machine host

Virtualization's flexibility makes it possible to provision a virtual server in a matter of seconds, run it for exactly the time your project requires, and then shut it down. The resources released will become instantly available to other workloads. The usage density you can achieve lets you squeeze the greatest value from your hardware and makes it easy to generate experimental and sandboxed environments.

Major cloud providers like AWS have enormous server farms where hundreds of thousands of servers and disk drives are maintained along with the network cabling necessary to connect them. A well‐built virtualized environment could provide a virtual server using storage, memory, compute cycles, and network bandwidth collected from the most efficient mix of available sources it can find.