Slide title In CAPITALS 50 pt Slide subtitle 32 pt ANALYSING EFFECTS OF MALFUNCTIONS ON THE PERFORMANCE OF UMTS RADIO ACCESS NETWORKS Author: Antti Keintola Supervisor: Professor Sven-Gustav Häggman Instructor: M.Sc. Anna Merjamaa
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Agenda Background Research problem and objectives Methods Analysis Results Summary
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Background 3G networks are launched all over the world Technology is complex and still somewhat immature Not much experience in operating and troubleshooting the network Networks are not performing at the optimal level Faults and failures occur Troubleshooting is time consuming and finding the root cause is difficult Performance statistics is sometimes the only source of information available
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Research problem and objectives What faults contribute most on the performance degradation? Is it possible to find the cause of the failure by inspecting statistics? What actions should be performed to improve performance?
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Methods Counter data was collected from a live network –7 RNCs, 1906 Node Bs –28 day period –Only daily statistics, not hour or quarter data available Improvements made simultaneously Data analysed with spreadsheet tool –Laborious, slow –Need for dedicated tool
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Key Performance Indicators defined: –Call Completion Success Rate (CCSR) –Accessibility –Retainability CCSR = Accessibility * Retainability Methods
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Analysis – network The net degradation of performance were calculated and dominant days were identified Detailed analysis focused on days highlighted with red and yellow DayNet contrib. DayNet contrib. 13,50%153,85% 23,69%163,92% 32,91%174,08% 42,63%183,20% 53,29%192,76% 63,61%203,40% 74,14%213,99% 84,52%223,75% 93,67%233,91% 103,43%243,25% 112,80%253,41% 124,81%263,46% 133,36%273,33% 143,87%283,44%
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Analysis – accessibility Accessibility was studied first –Almost all accessibility failures related to RRC rejects
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Analysis – RRC at RNC level Six distinct faults were identified by analysing RNC RRC statistics
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Analysis – retainability Speech and packet drops were studied separately
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Analysis – Radio Access Bearers Drop rate analysis were done at RNC level –Basically RAB releases were investigated
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Results – accessibility failures Following defects were identified and their net contribution to the overall performance were calculated Transmission fault affects three out of seven RNCs, hence higher figure Software and hardware defects have small impact on the overall performance FaultNet contribution Parameter fault2,19% Software fault1,09% AAL2 fault1,62% Hardware failure 0,95% Cabling fault in core network 1,20% Transmission fault 4,22%
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Results – accessibility failures
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Results – retainability failures Speech calls Packet connections FaultNet contribution Software fault1,66% Hardware failure 1,25% Transmission fault 6,68% FaultNet contribution Design fault10,72% Parameter fault11,29%
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Results – retainability failures
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Summary Defects with significant impact on the network performance were identified Impacts on the performance were calculated Man-made defects have more severe impacts than software or hardware faults The reasons for man-made faults should be worked out Working instructions and procedures should be revised and improved Need for dedicated tool to analyse data from customer network
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Traffic properties
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Traffic properties
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Traffic properties
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola Traffic properties
Top right corner for field-mark, customer or partner logotypes. See Best practice for example. Slide title 40 pt Slide subtitle 24 pt Text 24 pt Bullets level pt © Ericsson AB 2005Antti Keintola