Data Analysis through SAP HANA

awfulcorrieAI and Robotics

Oct 29, 2013 (3 years and 7 months ago)

332 views




1




TEAM4Solutions













Data Analysis through SAP
HANA











ISMT E
-
200 Fall 2012

Greg Zheng

Hung Tran

Julio Silveira

Michael Chepkwony

Ryan Talabis





2

Contents

1.0
Executive Summary

................................
................................
................................
................................
................................
......................

4

1.1 Company
................................
................................
................................
................................
................................
................................
.........

4

1.2 Information Technology Supplier
................................
................................
................................
................................
...................

4

1.3 Business Goals

................................
................................
................................
................................
................................
............................

4

1.4 Propos
ed Solution:

................................
................................
................................
................................
................................
...................

4

Part 1
-

Business Requirement
................................
................................
................................
................................
................................
.......

6

2.1 Business context:

................................
................................
................................
................................
................................
......................

6

2.2
Required Functionality
................................
................................
................................
................................
................................
..........

9

2.3 Business Benefit Justification

................................
................................
................................
................................
.........................

10

2.4 Success Metrics

................................
................................
................................
................................
................................
.......................

11

Part 2: Technical Specification and Prototype

................................
................................
................................
................................
..

12

3.1 Architectural Approach
................................
................................
................................
................................
................................
......

12

3.1 Data Collection
................................
................................
................................
................................
................................
.........................

13

3.2 Data

Analysis

................................
................................
................................
................................
................................
............................

13

3.3 Business Intelligence

................................
................................
................................
................................
................................
...........

13

3.4 Mapping business requirements to solution:

................................
................................
................................
.......................

13

3.5 Software solution

................................
................................
................................
................................
................................
...................

15

3.6 User Interface Dashboard (Flex)

................................
................................
................................
................................
..................

17

3.7 R Embedded Software Environment

................................
................................
................................
................................
.........

17

3.8 The SAP HANA database:

................................
................................
................................
................................
................................
..

17

3.10 SAP HANA Appliance:
................................
................................
................................
................................
................................
.......

19

3.11 SAP HANA Replication:
................................
................................
................................
................................
................................
....

20

3.
12 Integration with existing enterprise applications

................................
................................
................................
..........

20

Part 3
-

Implementation Plan
................................
................................
................................
................................
................................
.......

21

4.1 Solution Development/Deployment:
................................
................................
................................
................................
.........

21

4.2
Risk Management:

................................
................................
................................
................................
................................
.................

24

4.3 Operational Readiness:

................................
................................
................................
................................
................................
......

25

4.3.1 Change Management

................................
................................
................................
................................
................................
..

25

4.3.2 Application Support and Hardware Maintenance
................................
................................
................................
....

26

4.3.3 User Enablement:

................................
................................
................................
................................
................................
.........

26

4.4 Success Metrics

................................
................................
................................
................................
................................
.......................

26

5.0 Ackn
owledgement

................................
................................
................................
................................
................................
......................

28

6.0 References:

................................
................................
................................
................................
................................
................................
......

29





3


Figures


Figure 1: As
-
Is Business Process:

................................
................................
................................
................................
.

8

Figure 2: To
-
Be Business Process

................................
................................
................................
................................

8

Figure 3: Business process and analytics solution

................................
................................
.........................

12

Figure 4: Software Architecture

................................
................................
................................
................................
.

16

Figure 5: HANA Studio

................................
................................
................................
................................
.....................

18

Figure 6: SAP HANA appliance

................................
................................
................................
................................
....

19

Figure 7: SAP HANA Replication jobs

................................
................................
................................
.....................

20

Figure 8: Project Plan
................................
................................
................................
................................
........................

22

Figure

9: Project Plan
................................
................................
................................
................................
........................

24


Appendices

Appendix 1: Solution demonstration:
................................
................................
................................
.....................

30







4

1.0
Executive Summary

1.1
Company

GLOCO is a multinational company based in Massachusetts, USA with product development, manufacturing plants and
distribution centers in 4 continents. GLOCO manufactures a wide array of medical equipment and provides technical
support services for the whole product line.

1.2
Information Technology Supplier

TEAM4Solutions (T4S) is an IT consulting company that specializes in
the
devel opment of analytics and data
management solutions
that

support and optimize business operations. T4S is a certified part ner of SAP, Cisco
,

HP
,

IBM
,

and
other

companies with exper
tise in a wide range of development frameworks and technologies.

1.3
Business Goals

Product quality and customer satisfaction are key
to

GLOCO’s business success
. As such,

its executive staff

realize
s

that
the following issues must be addressed
:

1.

High occu
rrence of unschedul ed maintenance due to incorrect forecast of
time to li ve (
TTL
)

of equipment and
parts.

2.

High mean
-
time for help desk and field technicians to resolve customer problems.

3.

Low visibility for sales and upper management on customer problems,
suggestions and requests.

4.

Low visibility for product
defects
, usability issues
,

and customer requests

in product development
.

5.

Lack of data analysis of customer complaints, suggestions and requests.

It is clear to GLOCO
management that they have the data to

mitigate the issues address
ed

above; however, the data is
de
-
centralized across vari ous systems, and therefore, don’t lend the data
to
correlation
or

analysis.
This leads to
challenges not only in terms of obtai ning access to the data, but on how to prese
nt the data in a way that enabl es GLOCO
to discover and gain meaningful insights.

1.4
Proposed Solution:

GLOCO
,

through TEAM4Solutions
,

has decided to initiate an analytics project to consolidate and analyze various
structured and unstructured data source
s from
select

GLOCO business units. The goal of the project is to
create

a
platform that would provi de GLOCO users with real time access

to

correlation and analysis of GLOCO dat a. The primary
areas that will be focused on for this project will be the follo
wing:

1.

Provi ding on
-
site technicians with real
-
time access to issue
s

and solution information to facilitate faster
resolution of product issues.

2.

Providing proactive analysis of user complaints in terms of product
issues
, patches,

and

recall
s

frequency.

3.

Provi ding a means to analyze unstructured data from business units such as complaints resolution,
maintenance reports, knowledge bases, product documentation
,

and quality assurance reports.

4.

Provi ding hel p desk with an easily searchabl e central repository o
f correlated information to assist with
product and issue inquiries.




5

5.

Providing visibility to various business units regarding customer complaints and issues
in order to

facilitate
better product development and enhancements.



The solution that will be
implemented is a custom analytics application that is founded on the following technologies:

1.

SAP HANA Appliance (in
-
memory database and analytics engine)

2.

SAP HANA Studio (data modeling and management tool)

3.

R analytics engine and text mining packages

4.

Java
-
b
ased middleware

5.

Flex based user interface (browser and mobile support)

6.

Crystal Reports Reporting

7.

Active Directory Integration


Each of the technol ogi es
listed
above

were

specifically chosen to facilitate the processing of large amounts
of
unstructured
dat a, correl at
e

and apply analytics algorithms

to data sets
, enable access to different devices such as
mobile, enforce security, and provide visualization and reporting capabilities.













6

Part 1
-

Business Requirement

2.1
Business context:

GLOCO’s

management has several use
-
cases where the lack of analysis of the available information related to
product problems has impacted business. The following table has use
-
cases and how information analysis could have
improved the decision

making process
:


Use Case (As
-
Is)

Improvement with Data Analysis (To
-
Be)

GLOCO is in a very competitive market with low profit
margins and any equipment recalls or parts replacement
can jeopardize the profitability of product lines. An
example is a recent large scale reca
ll for GLOCOS's
GL3000 Vital Signs Monitor that is sold in large scale for
use on EMS ambulances, started having intermittent
failures. The product error was caused by a malfunction
in the USB port where one of the electronic components
had to be replaced
by the brand used on GL2500. After
further analysis of the returned devices, the lab personnel
found out that the problem was happening when device
was exposed to low temperatures.

A better analysis of user complaints would have shown
the regions where the

problems were happening with
the GL3000 Vital Signs Monitor, and given a clue that it
was a temperature related problem. The additional
information could have reduced the size of the recall
and speed up lab investigations, resolution and
assembly line upd
ate.

GLOCO maintenance contracts guarantee a steady
revenue and excellent margin, but unscheduled
maintenance caused by equipment failure reduces the
margin with the need to re
-
route technical personnel as
well as parts/components replacement before the
estimat
ed TTL. Unscheduled maintenance also means
equipment down time, which is critical for medical
equipment in clinics and hospitals, as they normally do
not have backup.

GLOCO knows that most of unscheduled maintenance is
caused by operation error or wrong c
omponent TTL
estimates, but there is not a process in place for
complaints/support case analysis that could trigger
proactive maintenance.

Collection of complaint information including
unstructured data like text, correlation of data and
ranking customer
complaints by the potential severity
and probability of risk would help with proactive
maintenance.

Unscheduled maintenance/fixes are taking too long.
Maintenance time adds overall costs and equipment
down time. The main justification from technicians for

the
high maintenance time is lack of information. The access
to GLOCO’s system from remote locations is slow and it
is hard to find the information they need.

Technicians would be able to resolve problem in the
field much faster with real time access of c
orrelated
information from complaints resolution, maintenance
reports, knowledge bases, product documentation and
quality assurance reports.

GLOCO’s help desk and internal sales requires high level
of expertise. GLOCO sells medical equipment, supplies
and

services, and the help desk needs to be more
knowledgeable about medical terminology used by its
customers.

The high level of expertise translates on higher salaries
and GLOCO management is looking for ways to reduce
Help desk and
internal sales required level of expertise
could be downgraded if answer for requests could be
easily searchable from a central repository with
correlated information from complaints, suggestions,
inquires for product and services availability. Most of the

required information is not recorded by help desk or
sales representatives because it comes in an



7

personnel cost by 1
0
%.

unstructured format like text or voice. Saving and
analyzing/correlating unstructured data will make easy
for help desk and internal sales to support custome
r
requests.

Gloco’s Help Desk, Sales and Field technicians have
information about product problems, product and feature
suggestions and well as inquiries for products that
customers would be interested in buying. All this
information is not consolidate an
d correlated and end up
never reaching product development.



Correlation of complaint information as well as data from
Help Desk, internal sales and field technicians,

would
be key for development of new products
and

enhancement of existing product.

Ana
lysis of the existing information would be helpful in
prioritizing the company resources, e.g., what new
products or features would maximize investments.

GLOCO has high volume of customer service complaints
and the resolution time of trouble tickets are
taking too
long. GLOCO would like to reduce the average customer
resolution time
.


Help Desk would be able to resolve problems and
respond to inquiries much faster with real time access
of correlated information from complaints resolution,
maintenance rep
orts, knowledge bases, product
documentation and quality assurance reports

Sales and marketing teams have low visibility of
customer problems, complaints and suggestions. It has
no visibility on how product development correlates with
customer
information.

Sales would have full visibility of customer problems,
complaints and suggestions and how it correlates with
product development. The information will help new
sales, up
-
sale and maintenance contract negotiations.


The information required fo
r analysis is available in different GLOCO databases, documents and spreadsheets. Most of
the data is unstructured and not used to provide business information.


Figures 1 and 2 show
s

the business process
:

As
-
Is
,

and To
-
Be after the information is consolid
ated and available for
analysis:








8

Figure
1
: As
-
Is Business Process:







Figure
2
: To
-
Be Business Process





9

2.2
Required Functionality

Based on discussions with GLOCO management and analysis of the
business problems
,

the functionalities
identified as
requirements for the solution

are
:

Analytics, Data Management, User Interface, Performance, Security
,

and Deployment.
What follows is a breakdown of the specific functionalities for each area that addres
s or support the various business
problems and proposed process changes.


Analytics Functionality

1.

Solution
is
required to provide
a
list of possible root cause of product issues.

2.

Solution
is
required to support correlation and analysis of data provided by

field technicians
,

and provides a
list of recommended solutions for a product issue.

3.

Solution
is
required to provide correlation between complaints and onsite field technician information.

4.

Solution
is
required to provide real time analyses of product pro
blems to speed up the resolution of problems
via help desk and field technicians.

5.

Solution
is
required to provide most common

critical issues and danger probabilities.

6.

Solution
is
required to identify trends in products and product issues.

7.

Solution require
d to identify pattern on inquires and requests collected by sales, help desk and field
technicians, to be used by product development to identify the need for new product, new features or
enhancements.

8.

Solution

is

required to do fast correlation and analy
sis of unstructured data.

9.

Solution

is

required to process transactional and analytical workloads (OLAP and OLTP).

Data Management Functionality


1.

S
olution
is
required to have

integration capabilities for multiple types of data sources.

2.

Solution
is
required
to map, transform and load extracts from multiple sources, including structured (database,
spreadsheets) and unstructured data (free
-
form text).

3.

Solution
is
required to have integrated tool for data modeling.

4.

Solution
is
required to store and analyze large

volumes of data. Initial estimate is 3TB.

User Interface Features


1.

Solutions user interface required to be accessible via Mac, PCs and M
obile Devices.

2.

Solution
is
required to include reports with rich graphs.

3.

Solution
is
required to include executive dash
boards for data visualization.

Performance


1.

Solution
is
required to generate search results in real time.

2.

Solution
is
required to generate reports in almost real time.






10

Security


1.

Solution
is
required to include user authentication and authorization
controlled by GLOCO’s security directory
infrastructure.

2.

Solution

is

required
to maintain

an audit trail for activities conducted within the system.

Deployment


1.

Solution

is

required to support modular deployment with easy expansion.



2.3
Business Benefit
Justification


Project Implementation Budget

$2,250,000

Hardware and Appliances

$350,000

Software Licensing

$375,000

Software Development and Integration

$1,300,000

Vendor Consultation

$125,000

Staff Training

$ 100,000


The implementation of the HANA

project will reduce cost for the company in the following ways:

1.

Reduce mean
-
time for

on
-
site problem resolution
by
5
0%

for supported products. This will improve customer
satisfaction and enable first contact resolution. The cost per contact and
on
-
site appearance by field
engineers
will be drastically reduced.

2.

Reduce personnel costs
--
First

level customer representatives and internal sales
cost

reduc
tion duo to level

of
expertise need
ed
.

3.

35%
reduction
on unscheduled maintenance.

4.

Reduce
d

call time, allowing personnel cost reduction and increase
d

customer satisfaction

5.

Reduction of discovery time for defective products will expedite the process of identifying the root
-
cause of
the problem.

6.

Increase
i
n sales with the analytic reports of comp
laint management. (0.5% on 10% of the product line)

7.

Reduction of products recalls and patches.

8.

Provide critical information for product research and development.


Benefits from Implementation

$7,656,000

Reduce first level Help Desk cost (10%)


$536,000




11

Re
duce unscheduled maintenance (35
%)


$1,000,000

Reduce time f
or on
-
site problem resolution (5
0%)


$620,000

Reduce patches and recalls (from 3 per year to 1)


$3,000,000

Increase Sales (0.5% over 10% of product line)

$2,500,000

2.4
Success Metrics

Based on the proposed solution and proposed process changes, a list of suc
cess parameters was identified:



Business Impact Metric

Baseline

Goal (first year)

Reduce onsite resolution time for unscheduled
product issues

6 hours

3 hours

Reduce unscheduled
maintenance requests

1000 onsite cases

750 onsite cases

Increase number of product enhancements and
updates per year

1 major release per year
per product

2 major release per year per
product

Reduce number of product issues or support
request reported
(increase product quality)

3000 per year

2
000 per year

Reduce Help desk and Internal Sales cost


10 % reduction

Increase in average customer satisfaction for
problem resolution rating

3 of 5 (neither satisfied or
dissatisfied)

4 of 5 (somewhat satisfied
)

Reduce products recalls and patches

3

1

System Operations Metrics

Baseline

Goal (first year)

Freshness of data in the system

N/A

Data import from data sources
every 30 minutes

Application and System Security

N/A

No critical or high risk items
identified in vulnerability
assessments

Reports


Saved reports executed in less
than 10 min.

Application Performance

N/A

All queries to complete within 10

sec

Implementation Metric

Baseline

Goal (first year)

Project Cost Variance

N/A

15%

Project
Schedule Variance

N/A

15%







12

Part 2: Technical Specification and Prototype

3.1
Architectural Approach

This section addresses the architectural approach to implementing the proposed solution. In addition, it will address the
requirements identified in
part 1 of the proposal. It also goes into detail on the software and hardware configuration
required to successfully deploy the solution.


Figure 3 represents the mapping between the business process and the proposed analytics solution. The main processes
illustrated are
:

data collection, data analysis
,

and business intelligence. These processes are mapped with the new
systems in the proposed analytics solution as well as the current system in place.



Figure
3
: Business process an
d analytics solution


Legend:


(1) Data from various BU* is imported to HANA database.

(2) Consolidated data is used for analytics.

(3) Data analysis is conducted using HANA and R.

(4) Data analysis results are transmitted to UI.

(5) Results are displayed

to BU users.

*Business Units (BU)
-

Help Desk, Service/Support, Sales, Marketing, Product Development, Quality Assurance




13

3.1
Data Collection

The process of data collection is supported through replication of the data stores identified in the Business Uni
ts (BU) of
GLOCO. The data from the BUs wi
ll be exported through SAP HANA

data management features. All data is then
federated in SAP HANA in memory database infrastructure.

3.2
Data Analysis

The process of data analysis is supported by SAP HANA
’s

built
-
in analytics capabilities and integration with the R
software environment. Consolidated BU data from the SAP HANA in memory database is processed through
the
analytics
algorithms and packages supported by SAP HANA and R
. This is done

to create, transform
,

and model data in order to
highlight useful information, suggest conclusions, and supporting decision making.

3.3
Business Intelligence

Information from the data analysis layer is presented to the user interface through a flex driven web server. User
inter
faces providing results and visualizations will allow users from various business units as well as employees who are
onsite with clients to access analys
i
s generated from the system and utilize them to gain insights and strategies based on
the topic or pro
blem they are currently working
on
.

3.4
Mapping business requirements to solution:

This section addresses the mapping of the requirements, i.e., functional, data management, user interface, performance,
security, and deployment, to the approach aimed at
satisfying these requirements.


Functional Requirements (From Part 1)

Approach (application, components, sites and services)

Correlation and Analyses

Solution must be able to provide the
following functionality: (i) Assist in early
detection of product
errors in production
line; (ii) Analyze and correlate data
provided by field technicians, and
recommend best possible solution; (iii)
Track on complaints in correlation with
onsite field technician information; (iv)
Provide most common critical issues by
p
roduct, and their failure probabilities; (v)
Identify trends in product malfunctions; and
(vi) Correlate customer enhancement
request to assist in prioritizing product
enhancements and updates.

For the analytics functional requirements, the proposed soluti
on will utilize the
HANA text engine and R libraries that are embedded in HANA and can be invoked
on SQL scripts.



The analytics layer will support various analytics algorithms such as association,
clustering and text mining algorithms. Specific R scripts

will be created and
customized to generate views and interfaces that will provide results that identify
root cause, recommended solutions, complaint correlation, danger probabilities,
trends and feature enhancements.

Solution must provide real time
analytics.

The proposed solution will utilize the SAP HANA in
-
memory database to facilitate
real time analytics on column based store that in
-
memory computing makes
possible.

Solution must have the capability to do fast
correlation and analysis of unstruc
tured
data (text and voice to text).

The proposed solution will utilize SAP HANA’s built in features that facilitate the
export and management of unstructured data. Several features that would facilitate
this are: (i) Upload of various unstructured formats

such as PDF,DOC,HTML and
Plain text to a BLOB field in a HANA DB; (ii) A Text Engine module supports various
text indexing and search abilities, such as exact search for words and phrases; (iii)
A Linguistic search feature that finds variations of words b
ased on linguistic rules;
(iv)

A Fuzzy search feature that allows for direct searching of various types of
unstructured data loaded to HANA; (v) Federated search feature that supports



14

searching across multiple tables and views; and (vi) SAP HANA supports f
ull
integration with R for statistical and text analysis.

Solution must have the capability to
process transactional and analytical
workloads using OLAP and OLTP.

The proposed solution will utilize SAP HANA to enable real
-
time online application
processi
ng (OLAP) analysis an online transaction processing (OLTP) due to both
row
-
based and column
-
based store engines in HANA.

Data Management Requirements

At a minimum, the solution must be able to
integrate with the enterprise databases,
email systems, text
documents and
transcribed voice data. In addition, it must
be able to extract structured (database,
spreadsheets, etc.) and unstructured data
(free
-
form text) from multiple sources.

The proposed solution will utilize SAP HANA to leverage its various data i
mport (and
export) capabilities. More specifically, these following features will be utilized: (i)
Export and Import feature in HANA Studio; (ii) Support of ODBC and JDBC to
connect to various data sources; (iii) in
-
line SQL support for importing various f
ree
-
form text formats to BLOB columns; and (iv) Federated search capabilities that
feeds queries out into other engines which in turn search multiple data sources.

Solution must be able to expand modularly

The proposed solution will be utilizing SAP HANA’
s scale out approach. This will
allow for the creation of multi
-
nodes networked together. This enables support for
larger SAP HANA memory sizes simply by adding compute nodes.

Solution must have integrated tool for data
modeling.

The proposed solution wil
l utilize SAP HANA Studio as the data modeling tool. The
SAP HANA studio is a collection of applications for the SAP HANA appliance
software. The Functionalities include: (i) Managing the SAP HANA database; (ii)
Creating and Managing User Authorizations; (
iii) Creating new or Modifying existing
Models of data in the SAP HANA database; and (iv) Accessing local or remote SAP
HANA databases.

Solution must be able to store and analyze
large volumes of data

The proposed solution will be able to process at least

4 terabytes of data. HANA
employs a scale
-
out architecture that allows it to expand the database beyond the
single
-
server boundary. The memory across all nodes in the cluster will be usable
as a single pool and thus technically, can be expanded to accomm
odate as much
data as required. In addition, The solution can expect a compression rate of up to
10 times due to column
-
based store that increases the likelihood of similar adjacent
data and large volume of text BLOB fields.

User Interface

Solutions user

interface must be accessible
via Mac, PCs and Mobile Devices (iOS and
Android).

The proposed solution will provide the following user interface options: (i) Utilize a
web based client that will be designed to support major browsers such as IE,
FireFox and

Safari; (ii) Utilize a mobile version of the interface (could be the same)
to allow for access to iOS and Android mobile devices; (iii) Utilize SAP HANAs built
-
in lightweight application server that can communicate via JSON which can be
subsequently read
by any web application scripting language

Solution must allow easy creation of reports
with rich graphs, to include executive
dashboards and data visualization.

The proposed solution will have the following features to generate graphs: (i)
Graphic
libraries supported by various web scripting languages (e.g. Flash, Java,
PHP) to generate graphics through the front end interface; (ii) Built
-
in visualization
capabilities of the analytics layer using R scripts; and (iii) Customized dashboards
based on t
he functionality needed per department via a custom frontend or through
the analytics layer.

Performance

Solution must be able to generate search
results in real
-
time and generate report
close to real
-
time.

The proposed solution will be leveraging SAP HA
NAs in
-
memory features to provide
real
-
time search and almost real
-
time reporting.

Security




15

Solution must be able to integrate into
GLOCO

s active directory (AD)
infrastructure.


The proposed solution will utilize AD integration with HANA DB. Security
and role
based permissions are managed by the Authorization Manager in HANA DB but
authentication can be delegated to an external provider such as the LDAP directory.

Solution must maintain an audit trail for
activities conducted within the system.

The pr
oposed solution will be utilizing the following approach for creating audit trails
for the system: (i) Utilize SAP
HANA’s

built
-
in audit policies and audit trail
capabilities. SAP HANA currently utilizes a syslog protocol to send and store audit
events to
a specified location; and (ii) Custom applications that will be created on
top of SAP HANA or the analytics layer will support auditing for access and
transactional events.

Solutions coding standards must follow
industry best practices (e.g. OWASP)

The fo
llowing security best practice approaches will be utilized: (i) In
-
house security
experts will be assigned to review the security architecture of the solution; (ii) All
custom application developed will strictly adhere to security best practices (e.g.
OWAS
P); and (iii) The proposed solution will include a vulnerability assessment and
a penetration testing in various development points in the project

Deployment

Solution must support modular deployment
with easy expansion.

The proposed solution will be
utilizing SAP HANA’s scale out approach. This
approach will allow the organization to create multi
-
nodes that are networked
together. This enables support for larger SAP HANA memory sizes simply by adding
computer nodes.

Solution must support fast deploym
ent of
hardware and software.

The solution will be utilizing SAP HANA appliances to facilitate ease of deployment.

3.5
Software solution

SAP HANA replication services will collect data from several enterprise data
sources.

The replication scripts will populate
data based on pre
-
defined data models and frequency. Scripts to load unstructured data from documents will use JDBC
to load the text to a BLOB column in a column table.

Most of the solution is based on analysis of
unstructured data loaded in memory and using the SAP HANA embedded text
engine that supports text search features, such as fuzzy or phrase search. To enhance the text analysis, the solution will
use R, which is embedded in SAP HANA.


Several algorithms pro
vided by the SAP HANA search engine and embedded R will be used in the solution, here are
some examples:

1.

The initial searches in the application will use the fuzzy search algorithm available in the HANA text engine,
calculating the fuzzy score for each str
ing. The score range from 0
-
10, and only records that meet the
minimum defined score will be displayed.

2.

Tree maps with text association will be done using R with apriori algorithm, with steps like:

(i)
Store the data
in a corpus (similar to an array)
;
(ii
)
Remove additional white space and convert the characters to lowercase
;
(iii)
Remove “stop” words (common words like a, the, are) using the Snowball library
; (iv)
Use a natural
language processor (NLP) to extract the phrases
; (v)
Use apriori algorithm to
retrieve the frequency of the
phrases
; and (vi)
Output the results





16

The user interface will be browser based and will be written in Flex. The Flex front end in turn will connect to the
middleware using BlazeDS. This process will allow flex and the back
-
end

to exchange messages in real
-
time.


The middleware will be written in Java and will be hosted on a Tomcat server. This will mainly be used as an interface to
call procedures in SAP HANA, but most of the application logic will be in HANA procedures with e
mbedded R, like
described previously. The solution will also have Crystal reports integrated with Flex to facilitate the display and creation

of reports.

All the data modeling, administration, and monitoring of the SAP HANA environment will be done via SAP

HANA studio, which is part of SAP HANA license. For user management, the HANA studio will be integrated to Microsoft
Active Directory.


All the Flex and Java development will be done by a 3rd party partner that has a strong relationship with SAP and has
i
mplemented similar solutions in other customers


During the software architecture design, several other options were considered, including the use of a commercial
software package integrated to SAP HANA like SpotFire or SAP Business Objects. The decision t
o develop the solution,
even though the above packages offer broader and richer features, was chosen because the alternatives would have
required extensive programming/customization to implement the text mining requirements. The packages also have
several

problems with integration and proper use of the SAP HANA calculation engine that can execute operations in
parallel.


Figure
4
: Software Architecture






17

3.6
User Interface Dashboard (Flex)

Adobe Flex is a development kit that is used to build Rich Internet applications (RIAs) that have functionality and features
of desktop application. Flex is chosen due its performance, product maturity, and robust tooling. Flex framework will be
used

to bu
ild the front
-
end of the dashboard for mobile and web users. Flex and the middleware interact with each other
using a remote data service called BlazeDS. BlazeDS transmit binary data over HTTP, which improve performance over
text
-
based protocols due to le
ss overhead.

3.7
R Embedded Software Environment

R is an open source programming language and software environment for statistical computing. Among the statistical
packages of R include robust data analysis capabilities including multiple text mining packages. These text mining
packages facilitates compl
ex text analysis and visualization functions of both structured and unstructured data.


The close integration between R and HANA allows R’s advanced data and text analysis capabilities to be leveraged in the
fast in
-
memory environment and data structures
of HANA. SAP HANA database allows R code to be processed in
-
line as
part of the overall query execution plan. This makes it possible for developers to embed actual R code (e.g. for text
analysis) in SQLScript (SAP HANA’s extension of SQL). SAP has specifi
cally created this integrated environment so
users of HANA can generate complex statistical and text based analysis while leveraging the benefits of in
-
memory
computing and storage.

3.8
The SAP HANA database:

SAP HANA (High
-
Performance Analytic Appliance)
is a data source independent in
-
memory database that combines SAP
software and certified hardware components into an appliance


SAP HANA is not just a standard database loaded in memory for fast access. In general, database management systems
are either go
od for transaction workload or analytical workload, but not both. When both are need, the workload is
normally separated in different databases (OLAP and OLTP) with data getting extracted from the transactional DB,
transformed and loaded in the data wareho
use database for OLAP.


SAP HANA has a hybrid structure with two engines within the same DBMS for:

1.

Row
-
based store. Similar to traditional OLTP databases that stores relational data in rows. The recommendation
is to use row store for tables with small numb
er of rows, with many selects and updates of a single records, with
columns containing two many distinct values or when aggregation and fast searches are not a requirement.

2.

Column
-
based store. Similar to data
-
warehouses databases that store relational data

in columns. Column store
should be used when calculations are executed in a single or few columns, when table has large number of
columns, when columns have few distinct values, when tables have large volume of rows and columns
processing is required (li
ke for aggregations).





18

User can specify whether the table is to be stored by column or row. SAP HANA column
-
based store adds other benefits
like:

1.

High data compression rate because a column
-
based store is more likely to have the same value in contiguous
me
mory. Data compression also increases speed with more data loaded to CPU cache. Another interesting point
is that HANA is aware of compression and uses it to speed up operations like aggregation and scan.

2.

Column
-
bases store makes parallel processing eas
ier. The columns relevant to a query can be divided in subsets
and operations executed in parallel. If different columns need to be searched or aggregated, it can be done in
difference processor core. Columns can also split in section to be processes by di
fferent cores.


SAP HANA has a calculation engine that allows calculation to be done in the database without moving the data to the
application layer for processing, and also provides a text search engine for text indexing and search capabilities, includin
g
exact search for words and phrases, linguistic and fuzzy searches.


A common problem with in memory database is how to persist the data in case of power loss. SAP HANA keeps a log in
non
-
volatile memory for all changes made in the data and every committe
d transaction creates an entry in the log. At the
same time SAP HANA stores changed pages in “save points” and they are saved in non
-
volatile store every 5 minutes.
The combination of log and asynchronous “save points” allows SAP HANA to recover the data i
n case of power loss and
meet the durability requirement of DBMS ACID
(
atomicity, consistency, isolation and durability) where the other three
requirements are not affected by the database being in
-
memory.

SAP HANA studio is part of the HANA solution and
contains a collection of tools for database administrators and
developers to manage data and monitor the HANA database. It runs on the Eclipse platform and has a uniform user
interface for all the tools. The main applications in HANA studio are:

(i)
Admini
stration console. It allows the monitoring
and administration of one or more HANA databases. It includes user management, backup and recovery, configuration
changes
, performance

information and troubleshooting
; (ii)
Information Modeler. It allows users to
create new and modify
existing models of data. It allows the creation of attribute, analytical and calculation views
; and (iii)
Lifecycle management.
It allows automated updates of the HANA software.


Figure
5
: HANA Studio






19

3.10
SAP HANA Appliance:

SAP is agnostic about hardware vendor but the HANA hardware solution must to be certified by SAP and the vendors that
have certified version are Cisco, IBM, HP, Fujitsu, Hitachi and NEC. SAP defines standard sizes that need to be matche
d
by the hardware vendors. A SAP HANA hardware configuration has the following components:

Server
: Intel Westmere EX with up to 8 CPUs and 80 cores.

Memory
: 128GB per CPU of RAM, 8 CPUs and 80 cores to a maximum of 1TB of RAM

Fast Log storage
: normally

Fusion
-
io ioDrive Duo. Same size as RAM

Dis
k storage, SAS direct storage
or network
storage depending on the configuration.


The SAP HANA appliance is limited to the certified hardware vendors, and our selection took in consideration price for
different s
izes of appliance (512GB, 1TB, 2TB, 4 TB, 8 TB)
, prices

for upgrades from 1 TB to 2 TB and to 8 TB, disaster
recovery plans, maintenance costs, monitoring and automation capabilities, storage solution.


In general hardware prices are compatible and we sele
cted Cisco as the appliance vendor because Cisco has no
competitive conflict with SAP, unlike HP and IBM, Gloco’s network is based on Cisco devices and a partnership is already
in place, and Cisco solution includes “Cisco Intelligent Automation for SAP HAN
A” that is a SAP certified tool for
monitoring and automation of operations and problem resolution.


Figure
6
: SAP HANA appliance








20

3.11
SAP HANA Replication:

Data from multiple data sources will be loaded in SAP HANA.
Replication can be manage directly from SAP HANA Studio
using SAP Landscape Transformation Replication (This is simple solution to retrieve the data from the non
-
HANA
databases) Supports ODBC, JDBC, MDX Scheduled, Manual, or Real Time Replication SAP HANA
supports flat file,
excel type files, text type files, etc., that can be used by HANA Studio for documentation.


Figure
7
: SAP HANA Replication jobs


3.12
Integration with existing enterprise applications

Integration of GLOCO’s

We
b Applications with TEAM4Solutions includes three key areas:

Data Replication
,
Authentication

and
User Interface
.

SAP HANA Data Replication Server

D
ata from multiple
data sources
, e.g. GLOCO BU Databases
and files
, will be
loaded in SAP HANA in
-
memory data
base with the use of SAP HANA Data Replication Server.

Single Sign
-
on

T
EAM4Solutions provides Single Sign
-
On capabilities that will enable GLOCO to integrate with chosen
authentication mechanism. Single Sign
-
On allows GLOCO users to seamlessly log into
TEAM4Solutions Server based
on a shared login mechanism and Id. Supported authentication mechanisms include Lightweight Directory Access
Protocol (LDAP), Active Directory, Java Authentication and Authorization Service (JAAS).

User Interface Themes/Skins

T
E
AM4Solutions application can be customized to match the look and feel of GLOCO
Web applications and allow for rebranding of user interface. This is accomplished through the use of CSS based themes,
which allow for a set of changes to be packaged together i
n one view.

HTTP APIs

A
ccessing to TEAM4Solutions Analytics reports and dashboards can be called via HTTP and embedded
into GLOCO Web applications using iFrames. There are a few different integration scenarios that can be accomplished
with following mechan
isms:

(i)
Using URL parameters (j_username, j_password) authentication can be skipped allowing
for the login screen to be skipped
; (ii)
Using URL with type parameter to indicate the type of report element to display
; and
(iii)
Using plain URL to load up on
ly a dashboard.

Web Services APIs

T
EAM4Sol utions provides SOAP and REST Web Services APIs allowing for different components
of the applications to be integrated into GLOCO Web applications. This includes the repository services, scheduling
services, domai
n services and administrative services. Web Services requests are authenticated using Spring Security
and can be configured to use HTTPS.

Metrics

A
pplication will
publish metrics

like response time, parallel access,
and number

of accesses.




21

Part
3

-

Impleme
ntation Plan


The success of the SAP HANA deployment depends on the effective planning of the project with all the appropriate
GLOCO stakeholders. This includes identifying business unit representatives, system administrators, developers,
database administ
rators, etc. The table below identifies all the GLOCO personnel that must partake on the project
deployment, along with their contribution to the project.


Role

Responsibility

Chief Information Officer (CIO)

Gain acceptance from GLOCO senior management
for the project, and approve its
funding.

Business Unit
Representatives

Identify the BU data to be replicated to HANA

Project Manager

Manage the project tasks, guard against scope
-
creep, provide status reports to
CIO, and ensure that the project is
completed within the identified budget.

Enterprise Architect

Assist with ensuring that the solution remains closely aligned with GLOCO’s IT
strategic goals

Business Analyst

Assist with the business work
-
flow of the application

Software Developer

Develop

the applications required for the project

Quality Assurance Engineer

Ensure that functional testing is performed on the solution prior to deployment to
production environment.

Security Architect

Perform a Security Risk Assessment on the application and
ensure that
vulnerabilities are fixed and/or ensure that senior management accepts the risk for
unmitigated vulnerabilities.

System Administrator

Ensure that the required hardware and operating systems are
built

to specification.
In addition, the administ
rator with assist with the implementation of Single Sign
-
On
with Microsoft Active Directory.

Network Engineer

Ensure that the application is networked to the GLOCO infrastructure.

Database Administrator

Identify the database tables to be replicated to
HANA, and assist with the
implementation


4.1
Solution Development/Deployment:

The scope of this project is limited to the following areas: (i
) develop a user interface (UI) that will be accessible via mobile
devices, e.g., tablets, android handsets, IPADs, etc., and traditional application using browsers. This UI will be used by
field technicians, helpdesk representatives, sales representatives
, product development and
upper management
; (ii)
integrate the system user’s Microsoft Active directory credentials to the application via Single Sign
-
On technology; (iii)
integrate the UI with a HANA back
-
end, and ensure that information is presented to t
he user in real
-
time; (iv) replicated
BU data to HANA database; and (v) implement a reporting feature into the application.

The UI Dashboard, UI reports, Java middleware , HANA SQLScripts and R scripts will be developed by a 3rd party
partner.

In order to
effectively develop the solution in an orderly manner, a project plan must be

created. The project plan identifies

th
e required deliverables

to successfully complete the project.





22

Figure
8
: Project Plan



The following table has
the list of deliverables expected per task in the project plan:


Task

Deliverable

Hana Appliance
(software/hardware)



Appliance specification and order



HANA Test/Dev environments plan



HANA Appliance installation and configuration plan



HANA Appliance
training plan

Crystal Reports



Crystal Reports order (Crystal will be embedded on FLEX for canned reports, so install
and configuration will be done by the development group.

Data Modeling



Data modeling schema

Test Plan



Solution integration test plan



Features test Plan

Hana Appliance
Installation/Configuration/Trai
ning



HANA appliance installed and configured



HANA training completed

Test Environment




Hana Dev/Test environments configured




23

Test Data for Development



Test data loaded in the dev/test
environment

Documentation repository
review



Documentation reorganization and common repository defined and loaded.

Data Federation phase 1



HANA replication scripts and jobs for Application Development phase 1



Load Data in the HANA test environment

Data
Federation phase 2



HANA replication scripts and jobs



Load Data in the Hana test environment

Application Development
Phase 1



Features, including development and QE:



UI framework



Middleware Framework



Hana SQLScript framework



Crystal Report integration



Initial use case from backlog like most common issues per type and model
with fuzzy search for issues




Feature test cases execution results

Demo/ Feedback Review
Phase 1



Phase 1 demo to stakeholders



Feedback triage and application phase 2 plan review

Application Development
Phase 2



Features, including development and QE:



Additional use cases from backlog like:



Search criteria including date, location and customer type



Drill down to root cause / most relevant solutions (heat map) for basic
criteria like

type and model.



Issues per customer analysis



Searches including various document types.



5 Reports



Login interface



Mobile interface



Feature test cases execution results

Demo/ Feedback Review
Phase 2



Phase 2 demo to stakeholders



Feedback triage and
application phase 3 plan review

Solution Acceptance Tests 1



Solution test cases execution results (features from phases 1 and 2)



Stakeholders signoff

Gloco Security Evaluation




Partial Security acceptance report and signoff

Application Development
Phase 3

I

Features, including development and QE:

a

Additional use cases from backlog like:

i

Additional drill down to root cause / most relevant solutions (h
eat
map) for advanced criteria.

ii

Identify failure probability per type and model

iii

Search any string and s
how documents per relevance with matched
information highlighted

iv

Advanced search algorithms

v

Drill down on problem per customer.

b

5 Report

II

Feature test cases execution results


Demo/ Feedback Review
Phase 3



Phase 3 demo to stakeholders



Feedback triage and
application phase 4 plan review

Application Development
Phase 4



Features, including development and QE:

a

Additional use cases from backlog like:



Drill down on issues per location



User management (authorization)




24



Geographic map (regional visualization) for

issues
-

cluster and heat
maps



Autocomplete on searches



Help desk can searches to get related terms to details



Heat map for medical terms



Reports



Publish application metrics.



Feature test cases execution results


Demo/ Feedback Review
Phase 4



Phase 4
demo to stakeholders



Feedback triage and application phase 3 plan review

Application Development
Final



Features added to backlog during demos/tests/reviews



Test cases updated with additional features



Feature test cases execution results


Demo/ Feedback
Review
-

Final



Final demo to stakeholders



Feedback triage and application phase 3 plan review

Help Desk Application
Integration




Integration with help desk application implemented

Field Technician Application
Update

1

Update field technician application
to allow data load in HANA.

Application Regression Tests




Regression test case
s execution results. It include
s

QE tests of all application features.

Solution Acceptance Tests 2



Solution acceptance test cases execution results (all features and
integration)



Stakeholders signoff

Gloco Security Evaluation




Final Security acceptance report and signoff

Solution Training



Class/Lab training and online training (recordings).

Go Production



Final project signoff



Updated solution specification



Source
code Escrow


Figure
9
: Project
Milestones


4.2
Risk Management:

The Risk Management process documents the risk mitigation strategy for the system. This is not to be confused with a
Risk Assessment, which primarily deals with
vulnerabilities to the system. The Risk Management Plan deals more with
business related risk associated with the system. The following are the business risks associated with the development
and deployment of the HANA solution.





25

Risk

Description

Mitigation

Schedule
overrun

Impact:
HIGH

Probability:
HIGH


Schedule overrun is a concern
due to the fact that the application
may not be delivered on time due
to delays, ineffective project
oversight, or resource limitation.

The plan for mitigating the risk is by defining a clear project schedule that
identifies all critical tasks, milestones and availability of resources. The
schedule will be communicated to all team members and stakeholders
including regular weekly status up
dates to ensure that all project tasks
are on track. The methodology will incorporate multiple iterations and
demo/feedback to ensure that all functionalities are delivered on
schedule. Any deviations in the schedule are recorded, escalated and
discussed
by the project team as required to determine action plans to
maintain the schedule. Day to day activities will also include a review of
tasks and color coding of tasks if issues encountered (e.g. red for tasks
that have issues). All said activities will be

closely coordinated with the
GLOCO PMO.

Cost overrun

Impact:
HIGH

Probability:
HIGH


Possible causes for cost overrun
for this project could be due to
poor definition of requirements or
scope, inaccurate estimates of
software/hardware/personnel
costs th
at could lead to
underfunding of the initial budget.
Cost overrun may also be a
product of a schedule overrun.

The plan for mitigating the risk is by a thorough and detailed assessment
of the business and IT requirements vetted by key stakeholders of
GLOCO
. This ensures that scope and features as established before
budgeting. A thorough analysis of hardware, software and personnel
costs will be conducted during the beginning of the project. All cost will
include estimates based on cost contingency principl
es.

Technical
Expertise

Impact:
MEDIUM

Probability:
LOW

The project utilizes a relatively
new (e.g. SAP HANA) or relatively
complex (e.g. R) technologies. As
with all new technologies, there is
a possibility that technical
resources might be scarce and
not
available in house at GLOCO.

The plan for mitigating the risk is by outsourcing to a Team 4 partner
development company with strong capabilities on foundational
technologies of the project such as Flex, SQL, R, Java and SAP HANA.
Additionally, all nec
essary legal contracts and escrow agreements will be
in place between the development company, Team 4 and GLOCO.

Coordination
with Business
Units

Impact:
MEDIUM

Probability:
LOW


A large part of the project is
replicating the existing data with
various BU

in GLOCO to HANA.
This involves coordinating with the
various BU representatives to
ensure that the HANA
development team is familiarized
with the specific data elements as
well as the source system where
the information will be extracted.
Lack of coordin
ation and
communication between the BU
representatives and the HANA
development team could
potentially lead to delays leading
to schedule and cost overruns.

The plan for mitigating the risk is to ensure that all relevant business
units are well represented

on the project. BU representatives who have
expertise on the source systems will be identified and will be part of the
project team during initiation of the project.

Targeted users of the new analytics system will also be identified and will
be included
during application demos as well as application user testing
in order to familiarize themselves with the new system as well as get
their input.


4.3
Operational Readiness:

4.3.1
Change Management

Once the system deployment is complete, the system will be

registered as an official GLOCO system in the Change
Management Board (CMB) of GLOCO. Meaning, all changes to the system must go through an approval process that is
voted on by the CMB. Examples of the changes that require authorizations from the CMB are:

(i) patch updates to any



26

component of the system, from hardware to software; (ii) enhancements to the systems front
-
end, middleware, back
-
end,
or operating system/hardware; and (iii) any type of maintenance to the system. In addition, all changes, regardl
ess of
whether change management was involved or not, must have a corresponding change order in the GLOCO ticketing
system.

4.3.2
Application Support and Hardware Maintenance

Application s
upport will be provided by T
4
S

through an application support and service contract with GLOCO. The
following support functions will be provided:

(i)
Patches and updates to the User Interface, Middleware, Replication
Scripts, and the HANA database
; (ii)
Addressing of any application tro
uble tickets pertaining to the components above
;
and (iii)
Monitoring and maintenance of critical integration and data replication scripts.

Any new enhancements and functionalities requested by GLOCO outside of the original requirements will be addressed
in
a case to case basis as part of a support contract adjustment.

Hardware maintenance will be handled through a support con
tract with the hardware vendor
.

All changes stemming from application support and hardware maintenance activities will go through
the Change
Management Board.

4.3.3
User Enablement:

Every development phase will have a demo to stakeholders. Feedbacks and enhancement requests will be included in the
backlog and addressed in the new development phases / iterations. The project plan in
cludes two solution acceptance
phases for stakeholder to validate and approve the solution.

Training will be addressed in the following manner: (i) the System
Administrators will

be trained on HANA appliance
installation and
maintenance to

effectively ad
minister the HANA infrastructure. In addition, the Database administrators
will be trained on SAP HANA Studio; (ii) In
-
class training sessions will be conducted for all the end users, i.e., Field
Technicians, Help Desk Representatives, Product Development
, Sales
Representatives and

Managers. (iii)
Video

tapings
of these courses will be made available to the systems users through the GLOCO internal website, in the event that users
may require refresher training.

Access to the system is tied into GLOCO’s
Microsoft Active Directory domain using Single Sign
-
On; however, access is
granted through the GLOCO Help Desk. Prior to receiving access, the requesting user must first have written
authorization from his or her manager.

4.4
Success Metrics

The success of the project will be determined by grading the success criteria identified in Part one of this document. In the

table below, the last column on the right represents that grading of the success criteria after one year of the system being
opera
tional.




27

Business Impact Metric

Goal

Method

Reduce onsite resolution time for
unscheduled product issues

3 hours

Report from existing
On
-
site

application that measures
time spent on issue resolution.

Reduce unscheduled maintenance
requests

750

onsite cases

Report from existing
On
-
site

application that measures
requests received and resolved.

Increase number of product
enhancements and updates per
year

2 major release per year per
product

Corporate newsletters.

Reduce number of product issues or
support request reported (increase
product quality)

2
000 per year

Report from existing Help Desk application that
measures support requests received and resolved.

Reduce Help Desk call time

15%

Report fro
m existing Help Desk application that
measures time spent over the phone.

Reduce Help desk and Internal
Sales cost

10% reduction

Report from existing Help Desk application that
measures the number of customers served per
employee.

Increase in average
customer
satisfaction rating for
problem/incident resolution

4 of 5 (somewhat satisfied)

Use of Customer Satisfaction Survey to measure
customer satisfaction.

Reduce products recalls and
patches

1 per year

Use of Product development database to measure
product recalls and patches.

System Operation Metrics



Freshness of data in the system

Data import from data
sources every 30 minutes

Reports generated from system log file.

Reports

Saved reports executed in
less than 10 min.

Metrics published by the
new application
.

Application Performance

All queries to complete
within 10 seconds

Metrics published by the new application.


Implementation metrics



Project Cost Variance

15% Cost Variance

Signoff document.

Project Schedule Variance

15% Schedule
Variance

Signoff document.







28

5.0 Acknowledgement


TEAM4Solutions would like to thank the tireless staff of ISMT E
-
200; more

specifically, Zoya Kinstler,
Takayuki Iida, Bob Wittstein,
Jeff Parker,
and
Alvaro Galindo

from SAP for making our Harvard Extensi
on
School

experience whole, as we complete the capstone of the Information

Technology, Information
Management Systems ALM degree.





29

6
.0
References:

Massimo Pezzini, Daniel Sholler, SAP Throws Down the Next
-
Generation Architecture Gauntlet With HANA,
Gartner

Research, 13 October 2011

Roxane Edjlali, Donald Feinberg, What CIOs Need to Know About In
-
Memory Database Management
Systems, Gartner Research, 8 September 2011

Roxane Edjlali, Donald Feinberg, Who's Who in In
-
Memory DBMSs, Gartner Research, 10 September

2012

SAP market place website and public web sites: h
ttp://www.sap.com/solutions/technology/in
-
memory
-
computing
-
platform/hana/overview/index.epx

http://www.sap.com/asset/index.epx?id=ca94e007
-
8139
-
4d17
-
b380
-
896aad6f0770

SAP documentation
-

SAP HANA Database for Next
-
Generation Business Applications and Real
-
Ti
me
Analytics
.


http://www.experiencesaphana.com
.


http://www.sap.com/solutions/technology/in
-
memory
-
computing
-
platform/hana
.


www.redbooks.
ibm
.com/abstracts/redp4814.html
.


http://en.w
ikipedia.org/wiki/SAP_HANA

SAP Hana developer’s guide:
http://help.sap.com/hana/hana_dev_en.pdf







30


Appendix 1:
Solution demonstration:


Step 1:

An on
-
site field technician was called by a client hospital to fix a much used but relatively old
ultrasound machine, model number Medison R6, in the OB
-
GYNE department. When the technician arrived,
the only indication of the problem is a cryptic flashing

“ERROR CODE 50” text displayed in the screen of the
machine. Previously, the field technician would have had to pour over various product manuals, wikis or even
call an overseas product support unit in order to figure out what was wrong with the machine
but instead, the
technician takes out his mobile device and connects to the new GLOCO analytics application.



Step 2:

Once logged in, the technician simply types in “ERROR CODE 50” in the search box and within
seconds, the technician is presented with a
consolidated view from various sources of what “ERROR CODE
50” is and what are the most likely root cause of the issue.




Step 3:

Even though the technician misspelled the word “error”, the search still came up with results; the
HANA database still searches through the unstructured data using the fuzzy search method. The technician is
then able to drill down to review each root caus
e and is even presented with the most common solutions to
resolve the issue. The technician then utilizes the knowledge and insights he has gained from his research and
fixes the problem fast and efficiently.

err


Step 4:

The technician clicks on “Medison
” inside of the “Manufacturer” panel. The application filters the results
to list only Medison products.





31


Step 5:

The technician then clicks on one of the previous case to see more details about “error code 50”.


Step 6:

Instead of going through all the

previous cases to see all of the resolutions, the technician drags the
“Key Phrases” panel and dropped it into the “Results”
panel. The

heat map in the “Key Phrases” panel display
all the associated phrases from the “Event Summary”. The size of the square

corresponds to the amount of
time the phrase was entered in the “Event Summary”. According to the heat map, the most frequent phrase is
“replaced power supply”. The second most frequent is “rebooted system”. Bob can try to reboot system before
replacing
the power supply to fix the error code 50.






32




Step 7:







33









34