###BeginCaseStudy###
Case Study: 2
Contoso, Ltd
Background
Contoso, Ltd. is developing a patient monitoring solution for a hospital. The solution consists
of an Azure website and a set of mobile applications that health care providers use to monitor
patients remotely.
Monitoring devices that run the embedded version of Windows will be attached to patients.
The devices will collect information from patients and will transmit real-time continuous data
to a service that runs on Azure. The service collects and distributes data. The data that the
service provides must be accessible by the website and by the mobile applications.
Business Requirements
Patients
All patient data must be stored securely on Azure. Data security must meet or exceed Health
Insurance Portability and Accountability Act of 1996 (HIPAA) standards in the United States
and must meet or exceed ISO/ICE 27002 data security standards in the rest of the world.
Contractors
Third-party contractors will develop the mobile applications. All contractors must develop
the applications by using virtual machines (VMs) that are hosted on Azure. Only authorized
contractors and authorized IP addresses are permitted to access the VMs. The contractors can
use Near Field Communication (NFC) tags to launch Remote Desktop (RD) connections to
the VMs from NFC-enabled devices. For testing purposes, contractors must be able to run
multiple instances of mobile applications within the VMs.
Data Collection and Distribution Service
The service must monitor the patient data and send out alerts to health care providers when
specific conditions are detected. The service must send the alerts to mobile applications and
to the website in real time so that doctors, nurses, and caregivers can attend to the patient.
Partner organizations and diagnostic laboratories must be able to securely access the data and
the website from remote locations.
Current Issues
A partner that is testing a prototype of the website reports that after signing in to the website,
the partner is redirected to the settings page instead of to the home page.
The data from the patient devices is slow to appear on the website and does not always
appear. All patient devices online have active connections to the data collection service.
Technical Requirements
Contractors
All contractors will use virtual machines that are initially configured as size A3. Contractors
must sign in to the assigned VM by using IP addresses from a list of preapproved addresses.
Data Collection and Distribution Service
• The service runs Node.js in a worker role.
• The service must use at least 2048-bit encryption and must use port 8888.
• All patient information must be encrypted and stored by using a NoSQL data store.
• Data must be stored and retrieved securely by using RESTful endpoints.
• Data must NOT be stored within a virtual machine.
All deployed services must send an alert email to [email protected] when any of the
following conditions is met:
• The CPU Percentage metric is at or above 85 percent for at least 10 minutes.
• The Network In metric is at or above 2 KB for at least 10 minutes.
• The Network Out metric is at or above 2 KB for at least 10 minutes.
• The Disk Write metric is at or above 1 KB/sec for at least 30 minutes.
• The Disk Read metric is at or above 1 KB/sec for at least 30 minutes.
Website and Mobile Devices
The website must be secure and must be accessible only within the hospital’s physical
grounds. All mobile applications and websites must be responsive. All websites must produce
error logs that can be viewed remotely.
Virtual Machines
• All Azure instances must be deployed and tested on staging instances before they are
deployed to production instances.
• All deployed instances must scale up to the next available CPU instance at a CPU
usage threshold of 90 percent and scale down when the usage is below 10 percent.
Application Structure
Relevant portions of the application files are shown in the following code segments. Line
numbers in the code segments are included for reference only and include a two-character
prefix that denotes the specific file to which they belong.
###EndCaseStudy###
HOTSPOT
You configure alerts in Azure. The metrics shown in the following exhibit represent the average values for each
five-minute period.
To answer, make the appropriate selections in the answer area.
Answer: 
Correct.
Disk uses Storage and Network Bandwith. Although Network In/ingress is free.
Network Out generates alert. Network In has not impact on cost.
wrong
I want to correct myself. You are right. Network In has no impact on cost. The CPU _may_ have a cost impact, if it goes above the scale threshold and we scale out.
So the answer must be:
Network Out
Network In
Right.
According to this table https://msdn.microsoft.com/en-us/library/ff803372.aspx “Network in” should be the right answer.
CPU percentage has no direct cost impacts, but can cause scaling (“All deployed instances must scale up to the next available CPU instance at a CPU usage threshold of 90 percent and scale down when the usage is below 10 percent”), so indirectly can have cost impacts.
So? Who got it right?
In this scenario, we scale app at 85% of CPU, which we haven’t reached. So, for test purposes, thinking CPU might be best answer? Although, as stated, both ingress and egress do not increase cost.
whar is the correct answer?
Network Out should be correct as they gave 2KB as threshold and Network out crosses it