The communications technology journal since 1924 2014 • 4
Data without borders: an information
architecture for enterprise...
Data without borders: an
information architecture
for enterprises
Enterprise information needs to be relevant, available a...
Thesignificanceofthislackofcoher-
ence is most apparent when it comes to
sharing information – when, for exam-
ple, the sa...
applications, the information at both
the business and application levels
needs to be aligned and connected.
However, achi...
northboundintegrationtosupport
communicationwithanenterprise’s
semanticinformationmodels;and
southboundintegrationtosuppor...
The information architecture is
based on components to facilitate de-
coupling and late binding, which in
turn facilitates...
informationinanenterprise.Onlypro-
ducersassignedwiththeroleofinforma-
tionownercancreatenewinformation
objects and instan...
external data sources. Beyond basic
connection, this part of mediation
includes consolidation and complete
life cycle mana...
Munish Agarwal
is a senior specialist in
multimedia architecture
and chief implementation
architect for DUSS. He has
been ...
Ericsson
SE-164 83 Stockholm, Sweden
Phone: + 46 10 719 0000
ISSN 0014-0171
284 23-3224 | Uen
© Ericsson AB 2014
Upcoming SlideShare
Loading in …5
×

Ericsson Review: Data without borders: an information architecture for enterprises

1,198 views

Published on

Today’s information systems rely on mediation to use and share the data that is available within an enterprise. But this is changing. IT systems are moving away from traditional architectures, where information is hidden inside functions, toward an information-centric approach that separates information from functionality – data without borders. Breaking functionality and information apart creates a systems architecture that is flexible, supports business agility, and ultimately boosts the bottom line.

Published in: Technology, Business
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
1,198
On SlideShare
0
From Embeds
0
Number of Embeds
60
Actions
Shares
0
Downloads
44
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Ericsson Review: Data without borders: an information architecture for enterprises

  1. 1. The communications technology journal since 1924 2014 • 4 Data without borders: an information architecture for enterprises April 24, 2014
  2. 2. Data without borders: an information architecture for enterprises Enterprise information needs to be relevant, available and all parts of an organization need to share a common understanding of it. As the significance of enterprise information and business agility rises, an information architecture that can capitalize on the changing nature of information, how it is generated, and how it is consumed, is an important enabler for business evolution and growth. deliver services that provide answers to basic questions such as “Where am I?” to more complex queries, such as“Whereisthenearestgasstationthat sellsfreshsandwiches?” Informationvirtualization Traditional approaches to IT and sys- tems architecture are also changing – moving toward greater flexibility to enable business evolution. This change is in turn shifting the direction of sys- tem design toward greater virtualiza- tionandhigherlevelsofabstraction. For telco operators, service virtual- ization of networks and network func- tions is a key part of ongoing network transformation. Much of the work on software-defined networking (SDN) and Network Functions Virtualization (NFV) relates to the functional per- spective of virtualization. Somewhat less discussed, but just as important, is the management and virtualization of information. Similar to the way that SDN, through abstraction, separates control (management) from the data plane, information virtualization sep- aratesfunctionalityfrominformation. Natureofinformation Information is diverse. Many different typesofsystemsexistforstoring,using and modifying information, and many differentsemanticmodelsandformats are used to describe it. Even within the sameenterprise,informationisnotnec- essarilystructuredinaharmonizedway – there are for example, several ways to refertopayroll,whichcouldbedefined aswages,payrollorsalaries. subscriptiondatabases,andapplication interfaces.Theanalysisthatcreatesuse- ful information from this raw data can be monetized and offered as a service to other enterprises, but, for most busi- nesses,goodinformationmanagement is an essential component of customer careanddailyserviceprovision. The growth in machine-to-machine (M2M) communication is yet another contributor to the richness of informa- tion. M2M communication provides enterprises with remote access to real- time information accumulated by a range of mobile and static devices and sensors – information that can be used to optimize existing offerings or iden- tifyservicegaps. The availability of real-time loca- tion data has been one of the more sig- nificant game changers. This type of information offers developers a basis for building sophisticated applications with a higher level of subscriber inter- est. Map applications, for example, LARS ANGELIN, JATIN SURI, MUNISH AGARWAL AND AKSHAY MEHRA BOX A Terms and abbreviations BSS business support systems CDN content delivery network CDR Call Detail Record CEP complex event processing CIME canonical information model of the enterprise CRM customer relationship management M2M machine-to-machine NFV Network Functions Virtualization NoSQL not only Structured Query Language OSS operations support systems RAM random access memory RDBMS relational database management system SDN software-defined networking SOA service-oriented architecture TCO total cost of ownership TTM time to market For most enterprises and industries, information is an asset. It is the raw material used to develop product strategies, roadmaps and marketing campaigns – a key element of modern business practices. The way people and businesses create and share information has changed considerably. The mass-market adop- tion of the internet, social media, apps andgaming–aswellasthedevicesthat have made non-stop information shar- ing and gathering possible on a global scale – are just a few of the factors that havecontributedtothischange. Consumer insight, for example, is no longer limited to information gath- ered by a business in its interactions with customers. Today, knowledge about customers, their needs and their preferences can be assimilated from a wide variety of data sources, includ- ing networks, social media platforms, 2 ERICSSON REVIEW • APRIL 24, 2014 From mediation to information
  3. 3. Thesignificanceofthislackofcoher- ence is most apparent when it comes to sharing information – when, for exam- ple, the same information is used by multiple applications, or across differ- ent technologies, between different vendor solutions, or across the internal boundaries of an enterprise. The com- plexprotocolsdevelopedbythetelecom industry are testament to the constant needforinformationtransformation. Traditional ICT architectures tend to be function-centric. As shown in Figure 1A, they are made up of dis- creteapplicationsthatperformspecific network functions – each one with its own semantics, formats and storage capabilities. This architecture can be visualized as many small information islandsthathavelittleconcernforeach other. Unfortunately, one of the conse- quences of this approach is cost; inte- grating applications that need to share informationandseparatingsuboptimal storage utilization is time-consuming and can be quite complicated. Here are some general observations related to evolutionofsystemsarchitecture: thegrowingsignificanceofinformation isnotagoodfitwiththefunction-centric architectureoftraditionalsystems; thegreatertheneedforinformation sharing,thehighertheintegrationcosts; tomakeinformationreusableand sharableacrossthebordersofasystem, itneedstobepartofacommonand sharedinformationmodel; informationtransformationis undesirablebutunavoidable–itwillnot disappearcompletely; tobeofvalue,informationmustbe trustworthyandcorrect. From a software architecture perspec- tive, if information is already available, known and in a readable format, it is a fairly straightforward process to build anapplicationthatmanipulatesorana- lyzesthedatainsomewayifitisalready accessible.However,ifthestartingpoint for application development is to ful- fill a specific function without much knowledge about what information is available, then the process of accessing the right information and integrating itwiththeapplicationislikelytobedif- ficult and costly. With an information architecture in place, whatever infor- mation is available is known, and so no translationisrequired. Gap Enterprise business Traditional application landscape Traditional function Traditional function Traditional function Traditional function Traditional function Traditional function Traditional function Traditional function Traditional function Traditional function Traditional function Traditional function FIGURE 1A The enterprise information challenge Enterprise business BSS/OSS Common information and integration services Traditional BSS/OSS function Traditional BSS/OSS function Traditional BSS/OSS function Traditional BSS/OSS function Traditional BSS/OSS function Traditional BSS/OSS function Traditional BSS/OSS function Traditional BSS/OSS function Traditional BSS/OSS function Traditional BSS/OSS function Traditional BSS/OSS function Traditional BSS/OSS function FIGURE 1B Overcoming the enterprise information challenge 3 ERICSSON REVIEW • APRIL 24, 2014
  4. 4. applications, the information at both the business and application levels needs to be aligned and connected. However, achieving this is not a simple task,withanumberoffactorscontribut- ingtoitscomplexity,including: typicaltelcooperatorsmanageseveral hundred,eventhousandsof, applicationsandsystemsthatvaryin nature–includingOSS/BSS,networks, datacenters,contentsystemsand cabling; applications,whenviewedfroman informationperspective,areislandsthat arelikelytohavebeendeliveredby differentvendorsandoftensupport differenttechnologies; integrationisneededtogetapplications tocooperateandperformbusiness tasks–themoreapplicationsinvolvedin atask,themoreintegrationconnections areneeded,thegreaterthecostandthe longerthetimetocompletion; manyapplicationsprovidesimilarbut notexactlythesamefunctionality,they mayusethesameorrelatedinformation –thissituationcreatessynchronization andinformationmastershipissues; buildingnewfunctionsontopofmultiple applicationsthatusedifferent informationmodelsisachallenge;firsta commoninformationmodelisneededto translateinformationfromoneformatto anothersothatitcanthenbeshared; applicationsarecreatedandterminated inlinewithevolvingbusinessneeds; informationandfunctionschange independentlyofeachother;and thetrustworthiness,correctnessand relevanceaspectsofinformationalways needtobemaintained. Informationmediation As Figure 2 shows, information origi- nates from many different sources and is consumed by a wide variety of users (people, applications, domains and sys- tems) that can be either internal or external. If information at the applica- tionlevelisfederatedfrommanydiffer- entsourcesittendstobemoreuseful. InsideOSS/BSS TakingOSS/BSSasanexample,theinte- grationpointsinclude: horizontalintegrationwithintheOSS/ BSStosupportcommunicationamong applicationsofdifferentgenerations, technologiesandvendors; Informationontheinside Enterprises typically have many inte- grationpoints: withpartners,customers,suppliersand regulatorsonanadvisory,commercialor technicallevel; amongenterprisesystems;forexample, back-officefinancialsystemsneedto shareinformationwithsupply,delivery andordering,aswellaswithHRand payroll;and withproductionsystems–suchas technology-orvendor-specificnetworks andcontentdeliverynetworks(CDNs). Generally speaking, an enterprise defines the information it needs to run a profitable business as well as the characteristics of that information (static, structured, event-driven, real- time, transactional or streamed). The spread of information across any given enterprise is often extensive, spanning manydifferentfunctionalareasinclud- ing marketing, ordering, strategy, HR, production, and finance. Similarly, the range of applications that use informa- tion is diverse; including for example, CRM, back-office support, payroll, and resourceplanning. The challenge comes in connecting business processes to support appli- cations and production or delivery systems – a challenge that is rooted in thefactthatrules,characteristics,infor- mation and procedures are not usually harmonizedwithbusinessprocesses. The consequences of this gap are reducedornoagility,lengthyTTM,long timetocompletion,andhighTCO,allof whichreducetheabilityofabusinessto evolve. A step toward overcoming this challenge,asillustratedinFigure 1B,is touseasinglefunctiontoserveallenter- prise applications with information – making that vital connection between informationandprocesses. Enterprises tend to be made up of a number of domains, each with their ownsetofinformationsemantics,goals and characteristics defined at both the businesslevelandapplicationlevel.The businesslevel,atthetopofthesemantic information model, includes the enter- prise vocabulary and concepts that describe the interactions between the peopleandsystemsinvolvedinbusiness processes. The application level is part of the internal value chain of an enter- prise. Activities at this level tend to be disjointed and are as such regarded as asetofindependentdomains,inwhich each application manages its informa- tion,tailoredtoitsparticularneeds. Tohelpmaketheconnectionbetween business processes and supporting Enterprise production environments Mediation/ transformation Enterprise data model Other enterprise systems PartnerDeployed OSS/BSS architecture Transformation Ericsson application level information model Other Other Other application level information model Enterprise application level information model Enterprise business concepts model FIGURE 2 Enterprise architecture 4 ERICSSON REVIEW • APRIL 24, 2014 From mediation to information
  5. 5. northboundintegrationtosupport communicationwithanenterprise’s semanticinformationmodels;and southboundintegrationtosupport communicationwithproduction domains. Towardaninformation architecture Thetransitiontoaninformation-­centric architecture is shown in Figure 3. On the left hand side, mediation plays a strong role and the majority of appli- cations have their own information model. While on the right hand side, themajorityofapplicationssatisfytheir information needs through a common informationmodel. The new model is built using emerg- ing software architectures and makes useofavailableinformation.Inthisway, enterprises can take advantage of new businessopportunitiesastheyarise. The foundation of the information- centric model is business-level informa- tion that is fully owned and managed by the enterprise, together with a set of applications that support business pro- cesses.Intraditionalarchitectures,infor- mationispiggybackedandhiddeninside applications. In the new model, systems recognizeinformation-relatedfunctions asfunctionsintheirownright. Inthenewmodel,thesetofinforma- tion functions are located in one infor- mationenvironment–theinformation tier – that serves the entire enterprise. The available information set, which ineffectisthecommonandsharedlan- guageusedbyapplicationsandbusiness usersisdefinedbytheenterpriseforthe application layer. In practice, the infor- mation set is a federation of existing informationsourcesandisreferredtoas thecanonicalinformationmodelofthe enterprise (CIME). The CIME includes manydifferentinformationtypes,such asstatic,structured,analytics,real-time data,andstreamed. Withfulllifecyclemanagement,the enterprise can control and maintain governance – from creation to termi- nation – for every information object in the CIME. Full life cycle manage- ment supports all information events including creation, publishing, access- ing, formatting, and storing, allowing theenterprisetoevolveitsinformation setasitsbusinessdevelops. The information tier offers infor- mation to users as a service. How and what information can be accessed is published in an information catalog, including any restrictions that may apply. These restrictions can also be set through the information tier, allowing enterprises to set the appropriate level of quality, integrity and security for individualpiecesofinformation. As enterprises move deeper into the information-centric model and take full control of their information, appli- cations become reduced to the level of information consumers and delegated creators. There are a number of bene- fits to building systems with an infor- mationtierandCIME: lowerintegrationcosts; smallerapplications–astheinformation architecturedeliversmanyofthe informationservices; fewerapplications–thesame applicationcanbeusedtoperformthe sametaskondifferentinformationsets. Forexample,B2BandB2Cinvoicingare typicallytwoseparatefunctionstoday; speedierintegrationwithexternal systemsandotherinformationsources; fasterintroductionofnewservicesand businessfunctions; flexibilityforbestbusinesssupport;and informationqualitycontrol. Servicesoftheinformationtier The information tier is built with soft- ware components that offer services in a service-oriented architecture (SOA) manner using a layered responsibility structure. This approach provides scal- ability, allows for introduction of new functions, and supports partitioning. The information tier is event-driven, and its behavior is controlled by enter- prise-defined rules and policies – sup- portingsimplifiedadaptiontochanging business needs. All management and governancefunctionsarecentralizedto gainbetteroverview,commonmanage- mentstyle,onemanagementconsole. Some of the characteristics and ser- vicesoftheinformationtierinclude: InformationasaService–allinforma- tionobjectsareofferedasaservicefrom acentralcatalog. Immutable objects – when informa- tionchanges,insteadofoverwritingthe existing stored object, a new instance iscreatedwithitsdistincttimestamp. Analytics – as all data passes through the information tier, it is the natural place for analytics and event process- ingfunctions. Two-way connection – to simplify the integration of information coming from, say, an external data source or a legacyapplication,theinformationtier includes a bidirectional mediation and connectionhandlingservice. Partitioning – is sometimes needed for legal,business,tenancyorperformance reasons. The information tier can be partitioned by information model, by storage,geographicallyorlogically. Storage – is a generic function that serves the entire CIME. Such as Time Role of mediation Application layer Information tier Information content A B E F G INF MED C D Application layer Information tier A B C F G E INF MED D FIGURE 3 Transition to information-centric systems 5 ERICSSON REVIEW • APRIL 24, 2014
  6. 6. The information architecture is based on components to facilitate de- coupling and late binding, which in turn facilitates configurable run-time characteristicsandreconfiguration. Informationmodels Meta information Every piece of information has an asso- ciated a set of metadata – time stamps, informationarchitecturefunctionality, languagetranslations,accessrightsand storage. Metadata plays a vital role in thecorrecthandlingofinformationser- vices,withitsownlifecycleandevenits ownmetadata. Each information object has rules associatedwithit,governingitsowner- shipandwhatactionscanbeperformed on it, such as create, publish, expose, access, store, and mediate. These rules are part of the system’s meta-informa- tion and are verified and approved by theinformationmanagement. Application-level information model The CIME helps to simplify application integration and ensures that applica- tion-level support exists. The CIME is usedforallinter-applicationcommuni- cation and mediation, and defines the commonandsharedsetofinformation. Itisfullycontrolled,lifecyclemanaged and governed by the enterprise. The information architecture can take the roleastheCIMEmasterorjustbeasup- porter,ifanothermasterexists. The CIME is an ontology, meaning thatalltherelationshipsofaninforma- tion object are included in the model and are rules-based. The relationships themselves are also defined as infor- mation objects and also adhere to the rules of the information architecture. Building the CIME in this way ensures that the connection between metadata andbusinessprocessesandtherelation- ships between information types can be maintained, as well as allowing for transformationsandsupportingdiffer- ent storage formats. Devising a CIME from scratch is both time-consuming and difficult. To simplify the process, a numberofblueprintversionsexistthat fitdifferenttypesofenterprises. Producers and consumers The information architecture provides servicestoproducersandconsumersof Data access and grid services management Metadata functionality Information model Non-real-time API Customer and partner management Information management Application layer Data grid framework Data engine framework Data vault Enterprise catalog Event management Analytics management Bidirectionalintegration andtransformationservices Data access and grid services Real-time API Non-real-time API Data engine services Real-time API Published IM Information accessTransaction support ESB distribution Reference and global ID Analytics and CEP engines Storage schemas engine Storage Information service registry Integration master/slave External data sources connection management Enterprise Master data management Reference and ID management Communication buses management Data access management Information model management Data transform management Grid and vault management Data engine services management Data persistence management Analytics and SEP engines management apps Other (legacy) FIGURE 4 Information architecture storage and persistence services for applicationsthatrequirelocaldata. Informationarchitecture Theproposedinformationarchitecture comprises three layers – a data vault, data engine and a data grid – that are governed by an information manage- ment function. The detailed architec- tureproposalisshowninFigure 4. The data vault offers different stor- agetypes,includingdisk,randomaccess memory (RAM) or virtual resources. The data engine contains a set of com- ponents that provide access to the data vault (storage), using a variety of tech- niques such as RDBMS, graph data- bases, or NoSQL. The information owner decides which type of database best fits their data access and persis- tency requirements. The data engine is responsibleforexecutinganalyticsand complexeventprocessing(CEP). The data grid is responsible for information creation, mediation and ensuring that the appropriate storage requirements are met. The data grid exposes information services to infor- mation architecture users. The appli- cations that consume and produce information sit on the top of the archi- tecture, or to the side of it. Each appli- cation places requirements on the grid layer for the information it needs, with the right characteristics and the appro- priatelevelofaccessibility. The information management func- tion is responsible for management, controlandgovernanceoftheinforma- tion architecture. It includes functions such as definition, creation, registra- tion, discovery, usage, archiving and decommissioning. This function is also responsible for the definition and managementofanalyticsandCEPalgo- rithms in conjunction with analytics andeventapplications. As data is no longer hard-coded by each application, but instead shared among multiple applications, the man- agement function is a vital element of theinformationarchitecturetoensure that data is consistent, available and robustatalltimes. 6 ERICSSON REVIEW • APRIL 24, 2014 From mediation to information
  7. 7. informationinanenterprise.Onlypro- ducersassignedwiththeroleofinforma- tionownercancreatenewinformation objects and instances of objects. This roleisassignedbytheinformationman- agementfunction,whichcanalsoadopt theownerroleitself. The information architecture holds the master information objects. Applications – even information own- ers–thatconsumespecificinformation objects must use the system for access to that object. All applications con- nected to the information architecture share common information through the same mechanism; all integration issues regarding external data sources arehidden. To ensure data consistency, infor- mation must be immutable. To meet this requirement, changes are imple- mentedbycreatingnewinstancesofan information object with a unique time stamp.Thehistoryofpreviousinstances iskept,andinthisway,themostrecent informationisalwaysavailablebutpast scenarioscanalsoberecreated. Storage The storage and access formats for an information object are determined by its owner. To cater for the varying information needs of different users, several storage and access formats can be assigned to the same information object. The information architecture performs the transformation from one storageformattoanotherandalsotakes care of storage management. There is alwaysaselectedstorageforthemaster ofeachinformationobject. Varying user needs are typically related to parameters such as seman- tic abstraction level, latency, integrity, rawformatsforreal-timeperformance, batch/volume,andspecialrequests. Different formats require different storageschemas,andthedifferenttech- nologiesareofferedasservicestoinfor- mationproducers. There are far more applications that useinformationthanthereareapplica- tions that create it. Many applications arepseudocreators;inotherwords,they recreate information for internal use. Applications that create information are responsible for storing it, and safe- guardingitsintegrityandaccessrights. Asystemwithoutmasterdataiscostlyto operate and develop, quality can be dif- ficult to maintain and TTM tends to be longerthannecessary. Mediation Traditionally, the term mediation refers to the area of collecting and pro- cessingCallDetailRecords(CDRs)from network elements for the purpose of billing.However,inthecontextofinfor- mation architecture, mediation refers tothebidirectionaltransformationand integration services that are essential for for communication with external datasources. The right-hand side of Figure 5 shows the set of functions that support the flow of information in and out of the system. This part of the informa- tion architecture shields applications on the inside from everything that is ontheoutside(sothattheyareagnostic from an information and responsibil- ity perspective). Typically, this is where an enterprise interfaces with legacy systems and external data sources, and communication is achieved through grid components that expose access to information objects. Typical functions include integration, protocol adap- tation, handling of services and data streams,aswellastransformation,and identitymapping. Mediation components can be selected, configured and bundled together to create mediation ser- vices. The mediation components are arrangedsothattheymodeltheflowof information in and out of the system. Likeallotherinformationinthesystem, the data carried over communication channels, also needs to be understood andtransformedintotheCIME. Canonical representation supports abstraction of an information object intoitsconstituentelements,datatypes and placement. These generated object definitionsarethensuitableforfurther processing through transformation components. Integration There is more to integration than simply creating connections to Non-real-time API Data engine services Real-time API Data access and grid services management Customer and partner management Information management Application layer Data grid framework Data engine framework Data vault Enterprise catalog Event management Analytics management apps Non-real-time API Data engine services Real-time API Analytics and CEP engines Storage schemas engine Storage External data sources connection management Enterprise Master data management Reference and ID management Communication buses management Data access management Information model management Data transform management Grid and vault management Data engine services management Data persistence management Analytics and SEP engines management Transformation components Mediation runtime - Routing - Enrichment - Normalization - Correlation - Aggregation - Filtering - Duplicate detection - Load distribution - Transaction management - Flow management - Session management - Event definitions - Schema/metadata - Encoding - Decoding - Real time - Streaming - Batch/file - Unidirectional - Bidirectional Structured data management Communication components Bidirectional mediation services Other (legacy) FIGURE 5 Mediation architecture 7 ERICSSON REVIEW • APRIL 24, 2014
  8. 8. external data sources. Beyond basic connection, this part of mediation includes consolidation and complete life cycle management of connections to the world outside the information architecture. Thecommunicationcomponentsare self-sufficient; they can manage and encapsulate the complete life cycle of communicationwithanexternalentity –includingtheabilitytoinitiateorwait for connections, apply interval-based polling,andretrievedataon-demand. Communication components encap- sulate the integration points and are responsible for ingesting data from external applications or data sources, and for external distribution of data. A variety of communication protocols can be used for this purpose, and they can be real-time or offline, streamed or batch-oriented, unidirectional or bidirectional – all of which needs to be managed by the communication components. Transformation Transformation functions operate on the object representation of incom- ing and outgoing information content. A given transformation component encapsulatesprogrammaticlogicto–in asinglewell-definedstep–completethe transformation and deliver the infor- mation object to its intended destina- tionand/ormapitintoitscorresponding placeholder(accordingtotheCIME).The primary aim is to be able to transform seemingly non-compliant data into a CIME-compliantframeworkortotrans- formdatatobecompliantwithanexter- nalprocess. Transformation is a two-way pro- cess that can be achieved in a number of ways – information mapping, pro- cedural translation and orchestration – depending on the complexity of the transformation needed. Information mapping is not, as might be expected, exactly one-to-one, but can be used to transform information objects that closely resemble the CIME represen- tation. Procedural translation is used when algorithmic logic is required to translate an information object into the CIME representation. Otherwise, orchestration can be used for transfor- mationsthatarenoteasilyencapsulated intoasinglestepexecution,suchas: handlingcardinalitydifferences–in situationswherethereisamany-to-one orone-to-manyrelationshipbetween externalinformationobjectsandobjects thatareinternaltotheinformation architecture; achievingtemporalalignmentor asynchronouscommunication–sothat out-of-banddatacanbecorrectly sequencedonthebasisoftimestamps (orsomeothercriteria); correlatinginformationreceivedfrom multipleinformationsourcesto constructsingleinformationobjects;and aggregatinginformationthatissplit acrossmultipleexternalinformation itemsintoacombinedinformation object,usingthresholdsbasedonatime windoworothercriteria,suchas summationvalues. Eventprocessingandanalytics Analytics routines and CEP are placed in the data engine layer as they per- form information services requested byauser. Typically,servicesofferedincludetra- ditional queries, real-time analysis, big data analytics, data streaming, pattern identificationandeventdetection. Implicationsfortraditional applications The current approach of information management, and hiding information insideapplications,needstobemodern- ized.Informationcontrolandassociated business logic need to be externalized from the application and put under the controloftheenterprise. Manydataqueriesandanalysisfunc- tions need information originating frommultipleapplicationsandinforma- tion sources. To ensure analysis is com- parable, it should be performed by an separate analysis service and not inter- nallybyapplications. Applications may cache information andareresponsiblefordeterminingthe structure of this information. Cached information is always a replica of the master,andapplicationsareresponsible forassuringtherelevanceofthecached information – providing updates when necessary. For (real-time) performance reasons, applications may use caching tech- niquestostoredatalocally,butthedata vaultistheonlypersistentstorage. Conclusions Tomaketheshiftfromtraditionalappli- cations acting as isolated islands to a fully life cycle managed information architecture takes time. However, the shift to an information-centric archi- tecture is an incremental one, and so both approaches will exist in parallel. Mediationplaysapivotalroleinsetting upaninformation-centricarchitecture, but once the majority of applications conform to the architecture, its signifi- cancewillslowlydecrease. An information-architecture model provides enterprises with centralized and shared set of information services andfunctions,makingthevitalconnec- tionbetweeninformationandbusiness processes–thekeytoflexibility. An information architecture is a single, shared and stable information environment that is trustworthy and can be used by all the applications of an enterprise. It separates functional- ityfrominformation,hidesintegration aspects, and is responsible for storage andpersistence. Thehigherlevelofabstractionthatan informationarchitectureprovidesgen- erates many benefits. For example, as all applications within the system use the same semantics and formats, inte- gration costs are lower. Connections with external systems are simpler, as there is only one internal representa- tiontotranslatetoandfrom.Withonly onemodelandonemaster,qualitymea- sures are easier to enforce, and secu- rityiseasiertoassure.Functionalitycan be reused on different data sets, which reduces the number of applications needed to conduct business. And appli- cations become smaller as common information services are provided cen- trally. Ultimately, a single information architecture offers flexibility, allowing an enterprise to make the most of busi- nessopportunitiesthatarise. The primary challenge for the enter- prise is to define and manage the CIME thatdefinesitsbusiness. In developing the information archi- tecture concept described in this arti- cle, Ericsson has focused on the telco industry. Information services tend to begeneric,andsotheconceptis,inthe- ory,applicabletoanyindustry. From mediation to information 8 ERICSSON REVIEW • APRIL 24, 2014 From mediation to information
  9. 9. Munish Agarwal is a senior specialist in multimedia architecture and chief implementation architect for DUSS. He has been working at Ericsson since 2004, primarily in the OSS/BSS area. He is currently the driver for the OSS/BSS implementation architecture and is the product owner for Next Generation Execution Environment. He holds a B.Tech.in material science from the Indian Institute of Technology, Kharagpur, India. Jatin Suri is a senior specialist in mediation solutions and a member of the BUSS architecture team. He has been working at Ericsson since 2005 in the area of BSS/OSS and is currently the deputy chief architect for revenue management (invoicing domain). He holds a B.Eng. in computer science from the University of Pune, India. To bring you the best of Ericsson’s research world, our employees have been writing articles for Ericsson Review – our communications technology journal – since 1924. Today, Ericsson Review articles have a two-to- five year perspective and our objective is to provide you with up-to-date insights on how things are shaping up for the Networked Society. Address : Ericsson SE-164 83 Stockholm, Sweden Phone: +46 8 719 00 00 Publishing: Ericsson Review articles and additional material are pub ished on: www ericsson.com/review. Use the RSS feed to stay informed of the latest updates. Ericsson Technology Insights All Ericsson Review articles are available on the Ericsson Technology Insights app available for Android and iOS devices. The ink for your device is on the Ericsson Review website:www.ericsson.com/ review. If you are viewing this digitally, you can: download from Google Play or download from the App Store Publisher: U f Ewaldsson Editorial board: Håkan Andersson, Hans Antvik, Ulrika Bergström, Joakim Cerwall, Deirdre P. Doyle, Dan Fahrman, Anita Frisell, Jonas Högberg, Patrik Jestin,Magnus Karlsson, Cenk Kirbas, Sara Kullman, Börje Lundwall, Hans Mickelsson, U f Olsson, Patrik Regårdh, Patrik Roséen and Gunnar Thrysin Editor: Deirdre P. Doyle deirdre.doyle@jgcommunication se Subeditors: Paul Eade, Nathan Hegedus and Ian Nicholson Art director and layout: Carola Pilarz Illustrations: Claes-Göran Andersson ISSN: 0014-0171 Volume: 91, 2014 Akshay Mehra is the unit manager for M-Commerce RD. Prior to holding this position he was the system manager for multi-mediation, supporting the sales team with mediation-based solutions. He has been with Ericsson since 2008, working in the OSS/BSS area. He holds a B.Tech. and an M.Tech. in microelectronics from the Indian Institute of Technology, Bombay, India. Lars Angelin is an expert in the technology area multimedia management at BUSS. Lars has more than 28 years of work experience in the areas of concept development, architecture and strategies within the telco and education industries. Lars joined Ericsson in 1996 as a research engineer, and in 2003 he moved to a position as concept developer for telco- near applications, initiating and driving activities, most of them related to M2M or the OSS/BSS area. He holds an ­ M.Sc. in engineering physics and a Tech. Licentiate in tele-traffic theory from Lund Institute. 9 ERICSSON REVIEW • APRIL 24, 2014
  10. 10. Ericsson SE-164 83 Stockholm, Sweden Phone: + 46 10 719 0000 ISSN 0014-0171 284 23-3224 | Uen © Ericsson AB 2014

×