15 / 54
W H I T E P A P E R
www.persistent.com
© 2017 Persistent Systems Ltd. All rights reserved.
Cloud platforms also provide
message queue services
(e.g. Azure Service Bus, Amazon SQS, Google Cloud
Pub/Sub) which allows decoupling your components. The service provides asynchronous operations which
enable flexible, brokered messaging between clients and servers, along with structured first-in-first-out (FIFO)
messaging and publish/subscribe capabilities. While these are mainly used in transactional systems, this service
could be leveraged for implementing analytics rules.
Providers also have started offering a variety of services to build applications (e.g.,
API
and
workflowmanagement
),
as well as
mobile services, media services,
and
cognitive services
.
Cloud platforms also provide
backup
and
archival services
to import and export data to less costly storage. They
can also be used to encrypt data and keep it secured until it is required to be accessed. They are generally
available as part of the cloud database services (e.g., Azure), but some providers propose separate services for
these tasks (e.g., Amazon has AWS Glacier as a service).
In a cloud platform, you also need
Resource management
for both platform and
application-level resources
, i.e.,
support services for creating, deploying and managing data management and application artifacts. Each cloud
platform does provide those also as a service e.g. Azure Resource Manager, AWS CloudFormation, Google
Cloud Deployment Manager. This helps create flexible templates that deploy a variety of Cloud Platform services,
including your cloud database / data warehouse, data movement and your BI application. Finally, platforms
allow to monitor resources running on the platform: to collect and track metrics, collect and monitor log files, set
alarms, and automatically react to changes in your resources (AWS CloudWatch, Google Stackdriver, Azure
ApplicationInsights).
5 A complete example
The customer is a large business process outsourcing and professional services company, serving both the
public and private sectors. They work across eight markets, among which the education market. The customer
has a technology solution to manage student, parental and staff information, and is used by a large proportion
of schools in the country. The current production release, referred hereunder as “the OLTP product” is an on
premise, client-server architecture solution based around Microsoft SQL Server with business logic handled by a
custom .NET Framework module. Reporting and Analytics is an important module add-on of the overall solution.
PSL was tasked to develop a cloud-based reporting and analytics solution for the forthcoming version of this
product. Both this analytics solution as well as the OLTP product will be moved to the cloud.
5.1 Requirements
We summarize in the table below the requirements for the customer’s cloud analytics implementation in terms of
the dimensions defined in the previous section.
Factor
Requirement
Comments
Knowledge about
queries
Queries are known: canned and ad-hoc queries.
Types of queries
Canned queries (filtering, joining for operational
reports; OLAP-style queries for analytical reports)
Ad-hoc queries using potentially every field of every
table and user-defined fields
Hundreds of reports, tens
of dashboards
15