Question 231: How do containers perform isolation?
A. They perform application layer isolation.
B. They perform isolation at all layers like a virtual machine does.
C. They perform isolation of the repository.
D. All of the above are correct.
Correct Answer: A. They perform application layer isolation.
Explanation: Containers perform isolation only at the application layer. This is unlike a virtual machine that can offer isolation for all layers. Repositories require appropriate controls to be put in place to restrict unauthorized access to the code and configuration files held within.
Question 232: Which of the following is the number one security priority for a cloud service provider?
A. Implementing SDN firewalls for customers
B. Isolating tenant access to pools of resources
C. Securing the network perimeter
D. Offering network monitoring capability to customers
Correct Answer: B. Isolating tenant access to pools of resources
Explanation: The top priority for providers is ensuring that they implement strong isolation capabilities. All of the other answers are possible priorities, but “isolating tenant access to pools of resources” is the best answer.
Question 233: Which of the following are examples of compute virtualization?
A. Containers
B. Cloud overlay networks
C. Software templates
D. Containers and software templates
Correct Answer: A. Containers
Explanation: Of the list presented, only containers can be considered as compute virtualization. Software templates are used to build an entire environment quickly. Although you could use these templates in infrastructure as code (IaC) to build or deploy containers and VMs, this is not considered a compute virtualization. A cloud overlay network enables a virtual network to span multiple physical networks.
Question 234: What is/are benefits of a virtual network compared to physical networks?
A. You can compartmentalize application stacks in their own isolated virtual networks, which increases security.
B. An entire virtual network can be managed from a single management plane.
C. Network filtering in a physical network is easier.
D. All of the above are true.
Correct Answer: A. You can compartmentalize application stacks in their own isolated virtual networks, which increases security.
Explanation: The only accurate answer listed is that virtual networks can be compartmentalized, and this can increase security; this is expensive, if not impossible, in a physical network. SDN can offer a single management plane for physical network appliances, and the “ease” of filtering is quite subjective. Filtering in a virtual network is different, but it may or may not be more difficult.
Question 235: Which of the following components in a container environment require access control and strong authentication?
A. Container runtime
B. Orchestration and scheduling system
C. Image repository
D. All of the above
Correct Answer: D. All of the above
Explanation: Yes, all of the above is the right choice this time. But wait! There’s a good story here that I’m including for those of you still with me. In February 2018, Tesla (the car company) was breached. Thankfully for Tesla, the attackers only wanted to use Tesla cloud resources for bitcoin mining. How was Tesla breached? Was it a zero-day attack? Was it advanced state-sponsored agents? Nope! Its container orchestration software (Kubernetes in this case) was accessible from the Internet and didn’t require a password to access it! Not only did this give the attackers the ability to launch their own containers, paid for courtesy of Tesla, but inside the Kubernetes system was a secrets area that had Amazon S3 keys stored in it. The keys were used to access nonpublic information from Tesla. Again, container security involves much more than just application security within a container.
Question 236: True or false: Machine learning algorithms are stagnant with limited use cases.
A. True
B. False
Correct Answer: B. False
Explanation: AI tools are often easily transferable and can be tweaked for enterprise specific parameters. Businesses can use similar machine learning outlines for network security, fraud detection, customer relationship management, personalized marketing and recommendation engines on websites.
Question 237: What is one way that businesses can benefit from machine learning and AI technologies without directly involving data scientists?
A. Embedding machine learning into chatbots and other types of applications.
B. Uncovering information in large data sets that can be sold to other companies.
C. Using machine learning to analyze budget constraints.
Correct Answer: A. Embedding machine learning into chatbots and other types of applications.
Explanation: The past few years have seen companies deploying customer service chatbots that utilize AI to converse with customers online or on the phone, often without people even knowing that they’re talking to a machine. Chatbots are often a small step into digital transformations, as they have proven ROI and are a relatively safe way to get into machine learning.
Question 238: How does deep learning differ from conventional machine learning?
A. Conventional machine learning uses neural networks modeled loosely after the human brain.
B. Deep learning algorithms can handle millions more data points and run with less supervision from data scientists post-production.
C. There are no real differences between the two — they are the same tool with different names.
Correct Answer: B. Deep learning algorithms can handle millions more data points and run with less supervision from data scientists post-production.
Explanation: Even as traditional machine learning grows in use, its deep learning offshoot is becoming more common among companies doing particularly complex analytics projects. This is because deep learning applications can analyze massive amounts of data in a largely self-directed mode, without data scientists needing to supervise the analytics process as much as they do for machine learning jobs.
Deep learning tools can also lend a hand to advanced analytics work involving unstructured data that requires subjective interpretation.
Question 239: Machine learning can be used on many types of data, structured and unstructured. All of the following are examples of effective uses of machine learning, except:
A. Personalizing marketing campaigns based on customer demographics and purchase history.
B. Detecting possible fraudulent activity in financial transaction data.
C. Examining IoT data to predict equipment issues before they arise.
D. Analyzing previous product revenue data to determine the cause of dropping sales.
Correct Answer: D. Analyzing previous product revenue data to determine the cause of dropping sales.
Explanation: Machine learning software is typically used in predictive analytics because the technology makes it easy to find patterns and relationships in data sets and apply that information to current trends. For example, customer purchase data and internet clickstream activity can be used to predict what individual online shoppers may be looking to buy, enabling a company to target ads to them accordingly.
In addition, anomalies in financial data can help corporate security teams detect fraudulent activity. And abnormal readings coming from IoT devices can help predict when a device needs to be repaired to prevent equipment failures. However, using past data to determine something that went wrong doesn’t require any predictive capabilities — in this case, basic business intelligence and analytics applications usually will suffice.
Question 240: True or false: Machine learning applications are too complex to run in the cloud.
A. True
B. False
Correct Answer: B. False
Explanation: Doing machine learning in the cloud provides potential advantages for companies looking to streamline their analytics deployments; scalability, cheaper storage and the ability to easily test analytical models are chief among the benefits.
Some companies do have concerns about cloud analytics, including security and data governance issues, but with some safeguarding and infrastructure changes, numerous vendors can offer safe cloud-based machine learning platforms.