- Количество слайдов: 20
Mobile Radio Network
Contents • Introduction and background • Literature review: • • Architecture of Mobile Radio Access Networks State of the art in management of mobile networks as defined by 3 GPP Performance management data Functionality scenarios in UTRAN • Methods of the practical study • Results: • • • Results (I/IX): Current PM and FM organisation and process Results (II/IX): Example on current PM and FM process (1/3) Results (III/IX): Example on current PM and FM process (2/3) Results (IV/IX): Example on current PM and FM process (3/3) Results (V/IX): Analysis of the current organisation and process Results (VI/IX): Problems in current PM and FM process, Interrelationship- and why-diagrams of the process problems • Results (VII/IX): Summary of the analysis • Results (VIII/IX): Solution for automated investigation • Results (IX/IX): Implementation of the solution • Conclusions of thesis • References
Key concepts • 3 GPP = Project that aims to develop GSM and UMTS specifications in cooperation with the vendors, operators and standardisation organisations. The acronym 3 GPP stands for Third Generation Partnership Project. • Fault management = Functions that enable the detection and location of failures in the network and scheduling of repairs. 3 GPP specifies the requirements of the concept. • Mobile radio access network = a network that provides wireless access to users through radio interface and allows the mobile users to move between coverage areas without losing connection, i. e. handover. • Performance management = Functions that enable the performance measurements of network services. 3 GPP specifies the requirements of the concept.
Introduction and background How to enhance the productivity of the UTRAN performance management investigations? • Research area: Mobile Radio Network Performance and Fault Management • Research questions: • What is a mobile radio access network and how it is managed? • What is the current performance management set-up in the organisation under study? • What is the organisation and communication structure? • What is the process to tackle performance problems in UTRAN? • What are the problems of the current set-up? • What could be solutions to the root-problems found from the current setup? • Scope: • Limited to European 3 G mobile radio network = UTRAN (UMTS Terrestrial Radio Access Network) • Other major mobile radio network technologies are: GERAN (GSM radio network), Wimax (WMAN) and Wi. Fi (WLAN)
Architecture of Mobile Radio Access Networks • General architecture • UE – User Equipment • • RAN – Radio Access Network • • • Currently the most popular mobile RANs are UTRAN and GERAN Other radio access technologies are LTE, Wi. MAX and Wi. Fi CN – Core Network • • Consists of Mobile Equipment (ME) and Subscriber Identity Module (SIM) for the end-user to access the mobile network All RANs are attached to a CN that provides switching and access to services in PSTN and any IP network OSS – Operations Support System • All parts of the mobile network may be managed by a centralised system • UTRAN architecture • Network elements • • RNC – Radio Network Controller Node B aka. BTS – Base Transceiver Station A – ATM transmission nodes Interfaces • • • Iu. CS: RNC to Circuit Switched Core Network (voice and video calls) Iu. PS: RNC to Packet Switched Core Network (data calls) Iur: RNC to RNC Iub: RNC to BTS Uu: BTS to UE O&M: OSS to any network element: RNC, BTS, ATM-nodes and CN elements (MSC, HLR, SGSN, GGSN etc. ))
State of the art in management of mobile networks as defined by 3 GPP • Network management areas relevant to RAN technical support • Performance Management (PM) • Keeps track on the network performance status and analyses the effects of configuration changes in the network [3 GPP TS 32. 101] • Bases on measurements that are continuously recorded in the network elements • Fault Management (FM) • Consists of fault detection, fault localisation, fault reporting, fault correction and fault repair [3 GPP TS 32. 111 -1] • Bases mainly on alarms and system logs that the network elements produce • Software management (SWM) • Covers software request management, installation, customer feedback and software fault management, i. e. detection of software faults and finding resolution to the problems. This duty is close to and overlapping with fault management (FM) [3 GPP TS 32. 101] • Configuration management (CM) • Controls the operational parameters of network elements [3 GPP TS 32. 600] • Process of applying the network management: 1. Performance monitored 2. Faults localised Depending on the type of failure: 3 a. Configuration changed or 3 b. Software defect(s) corrected 3 a. CM 1. PM 2. FM param or software? 4. Monitor the performance (step 1) 3 b. SWM
Functionality scenarios in UTRAN • Control plane, i. e. signaling, on RRC connection (Radio Resource Control) • • Major purpose: setup and release a call User plane, i. e. the traffic, on RAB connections (Radio Access Bearer): • Major purpose: define the Qo. S class of the call: • • • Conversational class, RT (Real Time), applications: CS voice and video calls Streaming class, RT, applications: CS streaming video Interactive class, NRT (Non RT), applications: PS (Packet Switched) web browsing Background class, NRT, applications: emails, MMS (Multimedia Messaging Service) Signaling scenarios: • MTC (Mobile Terminated Call) scenario 1. 2. 3. 4. 5. 6. 7. 8. • Paging: RNC sends an “RRC Paging Type 1” message to the Uu interface RRC connection setup: The paged UE responses by starting the radio control connection establishment procedure by (1. ) sending an “RRC Connection Request” message to RNC (“RRC Connection Setup Attempt” counter is updated). (2. ) RNC tries to allocate radio resources (BTS) and if successful, it responses with “RRC: Connection Setup” message (“RRC Connection Setup Complete” counter is updated). (3. ) Finally the UE responses with “RRC: Connection Setup Complete” message (“RRC Access Complete” counter is updated). Transaction reasoning: RNC and CN negotiate on the transaction type Authentication and Security procedure: UMTS subscriber and network authenticate each other, and other security mechanism are activated RAB setup for transaction: Actual communication resources for the transaction are allocated. Transaction: UE has an active user plane bearer connection across the whole UMTS network RAB release for transaction clearing: Network resources related to the transaction are released, i. e. all the RAB active connections for an UE are released RRC connection release: Radio control connection between the UE and the UTRAN is released Mobility (handover scenario): 1. 2. 3. Measurement: the UE sends a radio-link measurement report to the RNC Decision: the final decision to make a handover is done in RNC by the RRM handover control algorithms. Decision bases on the handover criteria and algorithm parameters Execution: handover signalling between e. g. UE and RNC, and radio resource allocation e. g. in BTS
Performance management data • Performance counters • UTRAN collects thousands of counters that measure the amount of specific events • E. g. RRC Setup Attempts, RRC Setup Completes, RRC Setup Attempt Failure RNC, RRC Setup Attempt Failure BTS etc. • KPI (Key Performance Indicator) Calculated most often from performance counters to relative %-values • Relative% KPIs are comparable between networks of different sizes, absolute values are not, because the amount of traffic varies • Form: KPI = (a formula of performance counters) Examples: • RRC_Acc% = “RRC access complete ratio” = “RRC Access Completes” / “RRC Setup Attempts” • CSSR, Call Setup Success Rate (voice call) = RRC_Acc% * (RAB_voice_attempts-RAB_voice_failures) / RAB_attempts • CCSR, Call Completion Success Rate (voice call) = (RAB_active_voice_failures) / (RAB_active_voice_failures + RAB_active_voice_succesful_completes)
Methods of the practical study • Based on UCD (User Centered Design) process and framework • Chronologically the practical study had three phases: I. Study and define the current process and organisation a. b. Study: focus group c. II. Study: interview Study: contextual enquiry Analyse the current set-up a. b. Analysis: affinity diagram c. Analysis: double teams d. Analysis: interrelationship diagram e. III. Analysis: brainstorming Analysis: why-analysis Develop an enhanced process a. Solution: brainstorming b. Solution: SWOT analysis c. Solution: UML diagrams
Results (I/IX): Current PM and FM organisation and process • Organisation-wise Technical support is the communicator between the local customer contact teams and product line R&D organisation. 1. 2. Technical support investigates and analyses the performance degradations and makes decisions to fix them with the co-operation of R&D. 3. • Local teams communicate the performance status of the customer networks to the technical support. R&D’s responsibility is to develop corrections to the system, if no other solution is effective. The process follows the three phases of the root-cause analysis methodology: • • Analysis (maps to FM [3 GPP]) • • Investigation (maps to PM [3 GPP]) Decision (maps to SWM and CM [3 GPP]) Each phase of the process has deliverables that are utilised in the later phases.
Results (II/IX): Example on current PM and FM process (1/3) 1. Get KPIs and failure counters for the required top object (i. e. RNC) • KPIs: Achieved by using a reporting tool that collects the needed counters from the OSS measurement database and calculates the KPI values based on the counters. By manual postprocessing the data, the graphical output 2. Find measurement periods where there is a dip in performance: • • Call setup performance: at 11 the CSSR KPI has had poor values. The phenomenon has been partly ongoing during the next hour Retainability: high drop call ratio at 16. Counter diagram verifies that the drop in CCSR is due to high number of RAB active failures. Failure counters:
Results (III/IX): Example on current PM and FM process (2/3) • 3. Get the KPIs and failure counters on BTS level. • It can be achieved using the same reporting tool than in the first phase. The output is extensive list of all the BTS under one RNC, all measurement periods and counters per each BTS. • 4. Find the network elements that are causing the performance dip. • After post-processing the data, the results are lists of BTS that are the main contributors to the performance dips
Results (IV/IX): Example on current PM and FM process (3/3) 5. Gather the system logs for those network elements that are main contributors of the RNC performance dip. • Achieved by connecting to the network element’s O&M unit either by manual command procedures or using a tool that automates the procedure. The log files are usually in binary format, so they need to be opened by a parser or converted to textual format before the analysis can take place. 6. Analyse the detailed data. • The format of the data is vendor specific, i. e. not defined in public specifications => no general guidance can be set for the analysis itself. • Highly dependant on the individual system specialists that can handle the versatile analysis and can produce reliable results The analysis can be in this context treated as a black box, which has the input of system data, i. e. logs, parameters, alarms, counters and KPIs, and output of set of root-causes for the occurred performance problem. • 7. Generate a solution to the root-cause. • Needs the presence of a skilled system specialist. Depending on the type of solution, finding a working solution might need trial and error approach. • Before applying the solution to a live network, it is tested in a test bed of the vendor. Some network operators have also test beds of their own, on which they verify the solutions, e. g. SW corrections, before they are installed to the live network.
Results (V/IX): Analysis of the current organisation and process • Main problems: • Problems in current organisation operation • 7. 2. 1 High travel costs • 7. 2. 2 Troubleshooting poorly controlled • Problems in current PM and FM process • 7. 3. 2 NE logs not available for performance dips • 7. 3. 3 Alarms not mapped to performance dips • 7. 3. 4 Configuration data not available for performance dips • 7. 3. 5 Internal failures not distinguished from external causes • 7. 3. 6 Investigation is time consuming
Results (VI/IX): Problems in current PM and FM process, Interrelationship- and why-diagrams of the process problems
Results (VII/IX): Summary of the analysis • Analysis set two general requirements for the solution: • Support fault management analysis conducted by system specialists. The solution should be able to collect relevant fault management (FM) data, i. e. NE logs, configuration data and alarms, for troubleshooting. The evaluation of the FM data relevance bases on the performance measurement data, which may be collected either from OSS or from RNC. • Support general reporting of performance conducted by operator and vendor performance management bodies. The solution should produce scalable reports of the performance measurement data. Reports should represent the performance data both on whole network and individual network element level down to the level of a single cell. Other statistical requirements are: timely aggregation and that the data can be averaged.
Results (VIII/IX): Solution for automated investigation INPUT: • “Connection to a live network”. The requirement of the developed solution is either a working remote or onsite connection to the network. This prevents limitations on from which specific parts of the network the data is gathered, i. e. the OSS, NEs or some other databases in the network. OUTPUT: • System log files and other detail data for the failures that have occurred in the live network. The root-cause analysis phase utilises this data to make decisions. • Overview reporting of the network performance that can be utilised in reporting the status of the network to company management and to customer, i. e. the network operator.
Results (IX/IX): Implementation of the solution • The distributed system consists of five separate applications: • RNC monitor • RNC static performance data fetcher • OSS data fetcher • Processor & Report (application) • Report (server)
Conclusions of thesis • Summary of thesis, Thesis studied practical problems of mobile radio network management: • Conclusion: UTRAN vendor technical support requires a distributed system of troubleshooting tools to enhance its troubleshooting processes • • Purpose of the troubleshooting tools is to enhance the performance investigation by automating gathering of the performance and other relevant network behaviour data for the time periods where network suffers from low performance The reasoning of the solution bases on • Current troubleshooting set-up study: • • • The analysis of the current set-up: • • currently the main problem is the inefficiency of the first, i. e. investigation, phase in the performance and fault management process. Generalisation of the results • Same principles are applicable to other radio network (e. g. GERAN) performance and fault management • Utilization of an OSS in data gathering makes the solution more portable to other radio network systems • • Organisation: vendor home base technical support that is a link between local teams, which are located by the operated networks, and the vendor R&D. During special occasions, e. g. a new product release or emergency situation in network, the organisation may adjust itself by transferring temporarily system specialist to work locally by the operated network. Process: The practical performance and fault management process consists of three phases: investigation, analysis and decision. Typically OSS uses relational SQL databases. Different radio networks have different performance indicators. Then the same tools may be used after modifying SQL-queries, which is a straightforward process Future work • • Scope was limited to investigation. Also the complex analysis-phase has demanding development needs. Technical support organization requires product-processes to manage the development and maintanance of the troubleshooting tools.
References • Standards and Technical Specifications • 3 GPP: GSM, 3 G and LTE • IEEE: Wi. Fi and Wi. MAX • Commercial material • Nokia Multiradio: http: //www. nokia. com/NOKIA_COM_1/Microsites/Nokia. World/Press/Multiradio_Press_Backgrounder. pdf • Cisco Wi. MAX: http: //www. cisco. com/en/US/netsol/ns 616/networking_solutions_customer_profile 0900 aecd 80334 a 23. html • Previous thesis’ • Kujala, Kimmo (2006) Expert System for Mobile Network Troubleshooting. Thesis. Diplomityö, TKK / Sähkö- ja tietoliikennetekniikan osasto, 2006. 72 p. • An attempt to build automated fault analysis tool system. The result in thesis was that automated analysis is still unreliable! • Utriainen, Juha (2004) UTRAN Operation System Security. Thesis. Diplomityö, TKK / Sähkö- ja tietoliikennetekniikan osasto, 2004. 64 p. • Gives a good overview on the UTRAN O&M (Operation and Maintenance) • Handbooks • Kaaranen, Heikki (2005) UMTS Networks – Architecture, Mobility and Services. Second Edition. JOHN WILEY & SONS. ISBN: 0470011033 • Nielsen, Jakob (1993) Usability Engineering. Boston: Academic Press, 1993.