You need to configure the distribution tracking application

###BeginCaseStudy###
Case Study: 2
Trey Research
Background
Overview
Trey Research conducts agricultural research and sells the results to the agriculture and food
industries. The company uses a combination of on-premises and third-party server clusters to
meet its storage needs. Trey Research has seasonal demands on its services, with up to 50
percent drops in data capacity and bandwidth demand during low-demand periods. They plan
to host their websites in an agile, cloud environment where the company can deploy and
remove its websites based on its business requirements rather than the requirements of the
hosting company.
A recent fire near the datacenter that Trey Research uses raises the management team’s
awareness of the vulnerability of hosting all of the company’s websites and data at any single
location. The management team is concerned about protecting its data from loss as a result of
a disaster.
Websites
Trey Research has a portfolio of 300 websites and associated background processes that are
currently hosted in a third-party datacenter. All of the websites are written in ASP.NET, and
the background processes use Windows Services. The hosting environment costs Trey
Research approximately S25 million in hosting and maintenance fees.
Infrastructure
Trey Research also has on-premises servers that run VMs to support line-of-business
applications. The company wants to migrate the line-of-business applications to the cloud,
one application at a time. The company is migrating most of its production VMs from an
aging VMWare ESXi farm to a Hyper-V cluster that runs on Windows Server 2012.
Applications
DistributionTracking
Trey Research has a web application named Distributiontracking. This application constantly
collects realtime data that tracks worldwide distribution points to customer retail sites. This
data is available to customers at all times.
The company wants to ensure that the distribution tracking data is stored at a location that is
geographically close to the customers who will be using the information. The system must
continue running in the event of VM failures without corrupting data. The system is
processor intensive and should be run in a multithreading environment.
HRApp
The company has a human resources (HR) application named HRApp that stores data in an
on-premises SQL Server database. The database must have at least two copies, but data to
support backups and business continuity must stay in Trey Research locations only. The data
must remain on-premises and cannot be stored in the cloud.
HRApp was written by a third party, and the code cannot be modified. The human resources
data is used by all business offices, and each office requires access to the entire database.

Users report that HRApp takes all night to generate the required payroll reports, and they
would like to reduce this time.
MetricsTracking
Trey Research has an application named MetricsTracking that is used to track analytics for
the DistributionTracking web application. The data MetricsTracking collects is not customerfacing. Data is stored on an on-premises SQL Server database, but this data should be moved
to the cloud. Employees at other locations access this data by using a remote desktop
connection to connect to the application, but latency issues degrade the functionality.
Trey Research wants a solution that allows remote employees to access metrics data without
using a remote desktop connection. MetricsTracking was written in-house, and the
development team is available to make modifications to the application if necessary.
However, the company wants to continue to use SQL Server for MetricsTracking.
Business Requirements
Business Continuity
You have the following requirements:
• Move all customer-facing data to the cloud.
• Web servers should be backed up to geographically separate locations,
• If one website becomes unavailable, customers should automatically be routed to
websites that are still operational.
• Data must be available regardless of the operational status of any particular website.
• The HRApp system must remain on-premises and must be backed up.
• The MetricsTracking data must be replicated so that it is locally available to all Trey
Research offices.
Auditing and Security
You have the following requirements:
• Both internal and external consumers should be able to access research results.
• Internal users should be able to access data by using their existing company
credentials without requiring multiple logins.
• Consumers should be able to access the service by using their Microsoft credentials.
• Applications written to access the data must be authenticated.
• Access and activity must be monitored and audited.
• Ensure the security and integrity of the data collected from the worldwide distribution
points for the distribution tracking application.
Storage and Processing
You have the following requirements:
• Provide real-time analysis of distribution tracking data by geographic location.
• Collect and store large datasets in real-time data for customer use.
• Locate the distribution tracking data as close to the central office as possible to
improve bandwidth.
• Co-locate the distribution tracking data as close to the customer as possible based on
the customer’s location.
• Distribution tracking data must be stored in the JSON format and indexed by
metadata that is stored in a SQL Server database.
• Data in the cloud must be stored in geographically separate locations, but kept with
the same political boundaries.
Technical Requirements

Migration
You have the following requirements:
• Deploy all websites to Azure.
• Replace on-premises and third-party physical server clusters with cloud-based
solutions.
• Optimize the speed for retrieving exiting JSON objects that contain the distribution
tracking data.
• Recommend strategies for partitioning data for load balancing.
Auditing and Security
You have the following requirements:
• Use Active Directory for internal and external authentication.
• Use OAuth for application authentication.
Business Continuity
You have the following requirements:
• Data must be backed up to separate geographic locations.
• Web servers must run concurrent versions of all websites in distinct geographic
locations.
• Use Azure to back up the on-premises MetricsTracking data.
• Use Azure virtual machines as a recovery platform for MetricsTracking and HRApp.
• Ensure that there is at least one additional on-premises recovery environment for the
HRApp.
###EndCaseStudy###

You need to configure the distribution tracking application.
What should you do?

A.
Map each role to a single upgrade domain to optimize resource utilization.

B.
Design all services as stateless services.

C.
Configure operations to queue when a role reaches its capacity.

D.
Configure multiple worker roles to run on each virtual machine.

Explanation:

* Scenario: distribution tracking application
The system is processor intensive and should be run in a multithreading environment.

Running multiple workers inside one Windows Azure Worker Role
http://mark.mymonster.nl/2013/01/29/running-multiple-workers-inside-one-windows-azureworker-role