SAS architecture for business analytics

WHITE PAPER

SAS A  Bsss Aas ®

SAS® delivers solution building blocks for empowering strategic business decisions within your existing technology environment

SAS® Architecture or BuSineSS AnAlyticS

Table of Contents Introduction .....................................................................................................1 ®

SAS Technical Reference Model for Business Analytics ............................. 2

Data Management .......................................................................................4 Analytics .....................................................................................................4 Reporting .....................................................................................................5 Compute Services ..........................................................................................6 Metadata Services ..........................................................................................7 Security Services ............................................................................................8 Integration Services.........................................................................................8 Management Services ....................................................................................9 Workow Services ..........................................................................................9

Five Styles of Business Analytics, One SAS® Architecture ..........................10

1) Classic Business Analytics ....................................................................10 2) Classic Business Analytics with Data Quality ........................................10 3) Business Analytics with Feedback Loops ..............................................11 4) Real-Time Business Analytics ...............................................................11 5) Business Activity Monitoring .................................................................12 One SAS® Architecture ............................................................................... 12 Conclusion ....................................................................................................13




Content or this paper, SAS® Architecture for Business Analytics, was provided by Diane Hatcher, Solutions Architect in the SAS Technology Practice, Cary, NC.




Introduction SAS has developed a business analytics architecture that can support many dierent architectural styles and business uses. At the core o the business analytics architecture are sourcing, discovery and sharing act ivities. Along with proper data collection and verifable data quality this leads to better analysis, which means more compelling inormation can be shared in a variety o ways with stakeholders, depending on the situational context. The details o what these activities involve vary across organizations, within organizations and even rom day to day. Only SAS provides an architecture exible enough to support organization-specifc requirements without sacrifcing perormance while providing a platorm that can evolve. Architecture PlAnning Business Analytics Activity

WHAT Functionality

WHY Value WHERE Stakeholders

HOW Technology

Source

Discover

Share

•Datacapture •Datacleansing •Datamanipulations •Masterdatamanagement

•Query •Statisticalmodels •Scoring •Visualization

•Publishing •Interaction •Integration •Reusability

Provideconsistentdata across the enterprise

Uncover opportunities

Distributeactionable intelligence

•ITdatamanagement •Operationalsystems

•Businessdecisionmakers •Decisionsupportsystems

•Businessinterfaces •Businessprocesses

•Dataaccesstools •Dataqualityalgorithms •Customertransformations •Jobowmanagement and scheduling

•SQLsupport •Analyticalmodeling •Customdevelopment •Interactivity

•Contentmanagement •Web-basedaccess •“Push”techniques •Servicesorientation

Figure 1. Architecture Planning Framework for Business Analytics

The diagram above (Figure 1) shows various attributes to consider when planning business analytics architecture. All elements and capabilities may not be required within each element, but the relevant ones should be linked and work together to provide a seamless solution. As the usage o business analytics grows over time, additional capabilities can be added. SAS provides a single ramework to meet all o your business analytics requirements, without the need to continually install new components. The SAS architecture provides a core set o capabilities that work together out o the box and can be easily extended to add more unctionality or integrated with other parts o your IT inrastructure.

1


SAS® Technical Reference Model for Business Analytics There are our key qualities that underlie and drive the SAS ®9 architecture or business analytics: • Scalability – the ability to meet perormance demands across the organization. Dierent stakeholders or interlinked business processes will have d ierent resource usage characteristics. The SAS architecture has the exibility to meet those diverse requirements. • Manageability – the ability to administer the SAS environment. As SAS becomes more ingrained in your organization, it becomes critical to manage and monitor usage to ensure optimal a vailability. SAS provides tools and techniques to manage the environment, either independently or within an existing environment. • Interoperability – the ability to use SAS capabilities rom outside the SAS environment. SAS provides an open architecture built to work with your existing IT inrastructure (Figure 2).

ENTERPRISE CLIENTS

Custom Applications

Enterprise Portals

SharePoint

Web-Based Application

StandardsBased Apps

Excel

PowerPoint

PDF

Outlook

Mobile

SAS

®

ENTERPRISE DATA AND SYSTEMS

Enterprise Applications SAP

Oracle

Peoplesoft

Siebel

Operational Data Teradata DB2 Oracle

Netezza HP Neoview AsterData Greenplum

Sybase SQL Server MySQL IMS-DL/1

JDBC ODBC OLEDB PC Files

OPERATING SYSTEMS

z/OS

AIX

HP-UX

Figure 2. Interoperability with Existing IT Infrastructure

2

Solaris

Windows

Linux


From supporting existing hardware and IT system standa rds to using virtually any database management system to integrating with enterprise communication channels, SAS has remained ocused on delivering decision support wherever and whenever needed. • Reliability – the ability to deliver results at the expected time. SAS delivers reliable results with powerul data access capabilities, industry-leading analytics and diverse tools or sharing content via multiple channels. SAS delivers reliable solutions, supporting multiple confgurations to provide highly available systems. Consistency o perormance in both results and delivery are integral to the SAS architecture.

Scalability Infrastructure Blocks

User Interfaces

Data Management Analytics y t i l i b a i l e R

Reporting

Framework Services

M a n a g e a b i l i t y

Interoperability

Figure 3. SAS®9 Technical Reference Model

Depicted above (Figure 3) is a high-level technical reerence model or the SAS ®9 architecture or business analytics. This represents a ta xonomy o the core entities, without stating a specifc relationship between them. The SAS architecture is made up o building blocks that encompass key characteristics o a service-oriented architecture. They are reusable services available across the suite o SAS tools and solutions to meet specifc unctionality requirements. Reuse allows SAS to provide service consistency and common integration points to ft within existing technology architecture. Framework services make up the application platorm that delivers common core capabilities across the breadth o the SAS architecture. Data management, analytics and reporting entities provide activity-specifc building blocks or sourcing, discovery and sharing. These blocks are a combination o targeted unctionality and user interaces designed to meet the needs o the specifc activity.

3


Data Management • Doyouhavetheneedtoaccessdataacrossmultipledatabasesandsystems? • Areyoucondentthatyourdataiscleanandconsistent? • Doyouknowwhatyourdatalookslike? The data management building block supports the business activity o sourcing. This could include prebuilt, high-perormance capabilities or connectivity, quality, extraction, transormation and loading, migration, synchronization and ederation o data. Data management means having a way to ully use all the data owing into the organization. This includes profling capabilities and the ability to incorporate data quality business rules across data sources and platorms. I you can automatically integrate data quality into data integration processes, you can ensure that data is current and accurate.

SAS® Data Management

Analytics • Doyouneedindustry-leadingstatisticalalgorithmsandvisualizationtechniques? • Canyouoptimizeandmanagetheportfolioofanalyticalmodelsacross the enterprise? The SAS Analytics building block supports discovery. The core o a business analytics architecture is the ramework support or an analytics engine, but it is how that analytics engine is suraced to the end user that determines its value. It is important that technologies are in place to support a robust analytics development process. This includes defning the business problem, developing and deploying the appropriate models and tracking perormance. Each o these steps requires due diligence and appropriate methodologies to deliver results to the organization. Analytics must be supported in a disciplined, managed environment to provide the maximum value.

SAS® Analytics

4


Reporting • Canallendusersaskquestionsandgetausableanswer? • Doyouhavetheabilitytopresenttheinformationintheappropriateformat? • Canyouusemultiplechannelstosharereports? Reporting should not be treated as a standalone activity. It is part o a seamless approach or creating and sharing intelligence. Reporting is more than querying. It includes properly presenting the inormation and sharing the inormation across the appropriate channels. The reporting building block is critical or the business processes o preparing and sharing inormation. This inormation must be presented in a way that is viable and delivered to the right person at the right time. Only when this is done can inormation become intelligence.

SAS® Reporting

5


SAS® Framework Services Framework services (Figure 4) represent the application platorm that delivers standard capabilities across the breadth o the SAS Business Analytics architecture. Data management, analytics and reporting address the components required or the core activities around sourcing, discovery and preparation, but they use a single inrastructure, representing the nervous system o the SAS architecture.

C o m p u t e

M e t a d a t a

S e c u r i t y

I n t e g r a t i o n

M a n a g e m e n t

W o r k  o w

Framework Services Figure 4. SAS Framework Services

These services are used by all SAS architecture building blocks to provide consistency o service and perormance.

comp Svs • DoestheengineyouareusinghandlemorethanSQLprocessing? • Canyoucreatecustomizedanalyticalmodelsbeyondwhat’sprovidedout of the box? • Cantheanalyticsengineaccessdatafromallexistingdatawarehouses and systems? At the heart o the SAS architecture are the compute services. Essentially, this is the Base SAS sotware that has been the engine o SAS Analytics or more than 30 years. The SAS engine remains the power o the platorm, executing queries and models to deliver intelligence. SAS has always been scalable vertically and horizontally, but we have also engineered SAS to be scalable across the enterprise or all types o users requiring both batch and interactive processing.

6


Feeding the engine is the SAS procedural language. This powerul language provides the ability to analyze data using hundreds o analytical algorithms – including simple queries, regressions, neural networks, time-based orecasting, sc oring and quantitative analytics. This is a huge advantage or users, because they are not limited to analytics delivered through a predefned user interace. Every organization is dierent, with dierent processes and dierent analytical needs. W ith user interaces that are amiliar to most organizations, SAS can support 80 percent o those needs. The remaining 20 percent can be customized to your organization to provideunmatchedanalytics.WithSASsoftware’smultivendorarchitecture(MVA), SAS code can be written once and run on any supported operating system. Analytics and data go hand in hand. The more data that is available, the better theanalyticalresults.OneofSASsoftware’scorestrengthshasalwaysbeendata access, and SAS is continuing to innovate with high-speed analytical data storage options, in-database processing support, and our commitment to continuous evolution o analytics techniques and user interaces.

Madaa Svs • Haveyouimplementedacentralmetadatalayeracrossthebreadthof analytic activities? • Doesthemetadataencompassbothtechnicalmetadataandbusinessmetadata? Metadata services connect all the business processes in the platorm – including resource management, security and sharing o content. Integrated metadata (inormation about data sources, how data is derived, business rules and access authorizations) is crucial or producing accurate, consistent inormation. SAS stores technical metadata and business metadata in an open, centralized and integrated repository. Data changes only need to be documented in one place. There are ewer systems to support and business users can count on high-quality inormation. A single version o the truth is available to all. Better use o sta time lowers the total cost o ownership or IT inrastructures. SAS integrated metadata also provides an auditable, repeatable and secure environment rom which to derive business intelligence. Content can be organized and controlled rom a central location, ensuring that enterprise users can only view the inormation they are allowed to see.

7


Sy Svs SAS security services provide additional measures to complement your existing security inrastructure and conorm to your e xisting standards or authentication, authorization and secure communications across the network. Userauthenticationcanbeonhost-basedauthenticationsystemsoruseLDAPor Active Directory mechanisms. Single sign-on is supported rom desktop applications running on Windows and through Web applications using trusted authentication. Metadata-based authorizations can be used to add an additional layer o access controls beyond what is supported on the fle system. Authorizations to SAScontrolled content and resources can be centrally managed via the metadata server. The implementation is exible to control both access to content as well as role-based access to application unctionality. For additional network security, SAS supports a number o mechanisms or encrypting network trafc. SAS/SECURE™ provides industry-standard encryption algorithms, including Triple-DES (168-bit encryption) and AES (256-bit encryption). These algorithms can be used to encrypt only credentials or all network trafc betweenSASclientsandservers.ForWebapplications,securesocketlayer(SSL) is supported.

iao Svs • Canbusinessanalyticsactivitiesandresultsbeintegratedwithexternal operating systems? • Cancustomapplicationsbedevelopedthatleverageanalyticalcapabilitiesto addressaspecicbusinessproblemforyourorganization? • Isitpossibletosurfaceanalyticalmodelstobeusedinexisting production environments? Integration services consist o technology components that acilitate integration o business analytics with other systems. The results o analytics can be used directly by individuals or can be automated into systematic decision processes. Integration services consist o technology components that acilitate SAS integration with other systems – either by invoking SAS algorithms or SAS calls to an external component. SAS algorithms can be suraced across SAS itsel, using the stored process interaces. This allows analytical modules to be reused across the SAS architecture in an encapsulated manner – delivering consistent results. These stored processes can also be defned as Web services, allowing external applications to invoke SAS or real-time scoring or other analyses.

8


External systems can be operational systems or custom applications. SAS provides a library o APIs to support access to SAS unctionality rom custom applications, and provides an Eclipse-based development environment to streamline creation o customapplications.LanguageinterfacesallowSAScodetobeeasilyintegrated withexternalsystems–whetherthey’rebasedonotherstandards(e.g.,PMML)or other platorms (e.g., Java or .Net).

Maam Svs Management services consist o technical components that monitor and optimize the use o business analytics across the enterprise environment. The SAS Business Analytics inrastructure is able to operate and integrate into existing IT enterprise system management acilities such as: • Systemmonitoringformonitoringthehealthoftheenvironment. • Eventmanagementforassessingthestatusofapplicationsvialogging, monitoring and implementing methods to correlate events, errors and warnings. • Systemadministrationfortheautomationofstart,stopandrestartsemantics. Management services include a technical component that provides the ability to log business analytics activity. The logs can be used to audit artiacts use, track perormance and monitor changes made to metadata security settings and other content.

Wokfow Svs Workow services defne the technical components supporting process management. These components are used by SAS solutions to provide the ability to defne business rules, trigger events and send notifcations when appropriate. Notifcations can be alerts via e-mail, invoking another process or triggering a Web service. Workow services allow or extension o workows or greater integration with enterprise business processes. It is possible to generate multiple-level approval and review processes. With standard APIs or accessing workow status or integration with business applications, analytics process ows become more manageable, visible and documented.

9


Five Styles of Business Analytics, One SAS® Architecture The SAS architecture is designed to be exible enough to support the fve styles o business analytics that are used by all organizations. Each style has its own set o architecture requirements. Multiple styles may be used at the same time within an organization. Understanding these styles can help you assess what architecture building blocks you should consider. For more details about these fve styles o business analytics, please read Architecture for Business Analytics: A Conceptual Viewpoint, available at www.sas.com/reg/wp/corp/17871.

1) Classic Business Analytics

Classic business analytics is the basic process o data sourcing and may include the creation o a data mart or warehouse, inormation discovery – via data exploration or predictive analytics techniques – and sharing generated reports or inormation at dierent levels o an organization. Traditional report-driven BI processes put inormation into the hands o users, leaving them to interpret the situational context and how it aects the business process. What is changing is how the inormation is distributed. Business stakeholders want inormation via e-mail, portals, dashboards and mobile devices. SAS Business Intelligence solutions can provide dynamic access to inormation, regardless o the interace, with the ability to refne questions, drill into more detail or visualize in a dierent manner.

2) Classic Business Analytics with Data Quality

In many ways this style is similar to classic business analytics except that it recognizes the need to cleanse and standardize the sourced data, so it integrates data quality into the data sourcing process. This could be critical to your organization to improve trust in the data or to meet regulatory requirements. SAS Enterprise Data Integration and DataFlux ® dPower® Studio provide prebuilt transormations and data quality algorithms to build data warehouses with standardized data. SAS can help you manage the entire data lie cycle to ensure that your data is clean, relevant and can be acted upon.

10


3) Business Analytics with Feedback Loops

This style supports cyclical business processes with a defned time window, where specifc inormation is needed to fnalize decisions. For example, you might want to provide specifc recommendations into a procurement workow. On a regular schedule, you may need to orecast sales to determine what you need to order to replenish your inventory. Data is extracted and c leansed using SAS Data Integration Studio, analyzed using advanced orecasting techniques with SAS Forecast Studio, and then specifc inormation on recommended purchase amounts is written back to the operational system via message queues. Procurement system users can then access the recommendations to better guide their decisions.

4) Real-Time Business Analytics

There are instances when it is unknown exactly when a specifc service might need to be executed to gather inormation because it is triggered when specifc behavior is observed. These situations beneft rom analytical enrichment by combining both contextual and historical data with very quick responses. This leads to a need or real-time business analytics. This style o business analytics reects the need to trigger analytics or data collection and delivery in real time rom an operational application. Contextual data is combined with existing historical data and analyzed. The results are sent back to the original touch point or a person to make a decision or or business rules to drive basic, automated decision making. One example is the real-time scoring o customers in a bank to see i they qualiy or a loan. SAS ® Enterprise Miner™ scoring models can be called via Web services, and the resulting score can be sent back via another Web service to the calling application to be used in the loan application process.

11


5) Business Activity Monitoring

Automated decision making is becoming the standard with technology-dependent operations, when human monitoring and decision making is generally not possible or timely. This requires monitoring key inormation, defning business rules and triggering alerts or other events that can drive downstream action. A services-oriented architecture is needed to provide sourcing, discovery and sharing to deploy business analytics directly into operational systems. SAS solutions, such as SAS Enterprise GRC, have begun to take advantage o workow capabilities in the SAS application platorm to support activity monitoring capabilities.

One SAS® Architecture The SAS architecture is unique because it supports all fve styles o business analytics rom one deployment. Because o the service-oriented approach, capabilities are encapsulated, extended and integrated to deliver the required business analytics. SAS architecture building blocks work together out o the box, supporting the ability to deploy some solution blocks frst and add additional modules later. The SAS application platorm – SAS ramework services – allows the inrastructure o the SAS architecture to be managed separately but delivered as part o any SAS solution. This provides exibility in delivering targeted building blocks to support your business analytics activities. There is no need to reinvent the unctionality provided by the ramework services when dep loying dierent SAS solutions.

12


Conclusion Regardless o the size, industry or goals o your organization, SAS provides an architecture or business analytics that can meet your needs today and into the uture. Key activities o sourcing the data, discovering what the data is telling you and sharing the inormation are supported with an integrated application platorm and exible building blocks. The SAS architecture building blocks are linked together to deliver consistent results. The success o your business analytics architecture depends on having: • Theappropriatebuildingblockstodeliverabusinessanalyticsarchitecture. • Theanalyticalcapabilitiesyouneedtosupportthebestdecisionmaking. • Ameanstointegrateanalyticactivitiestomaximizetheirvalueto your organization. The SAS Business Analytics inrastructure provides an e ective way to: • Managethegrowingappetiteforintelligence. • GaingreaterROIfromyourexistingITinvestments. • Supportsustainablegrowththroughinnovativeuseoftechnologyandinformation.

13

A

SAS is i. Wod hadqas

+1 919 677 8000

To contact your local SAS ofce, please visit: www.sas.m/fs SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. ® indicates USA registration. Other brand and product names are trademarks of their respective companies. Copyright © 2010, SAS Institute Inc. All rights reserved. 104763_S65253.1110

SAS architecture for business analytics

Recommend Documents