English

advertisement
FEEDBACK ON THE BOUYGUES TELECOM
EXPERIENCE WITH APPDYNAMICS
BY
CYRIL VANLATHEM
OSS FAI DEVELOPER
C. VANLATHEM - PAUG June 2013
Let’s protect the environment together. Don’t print this presentation unless it’s absolutely necessary..
AGENDA
1-
Bouygues Telecom
2-
Why an APM on the OSS?
3-
Choosing the APM from AppDynamics
4-
AppDynamics at Bouygues Telecom
5-
Specific use case
C. VANLATHEM – PAUG JUNE 2013
2
KEY FIGURES

C. VANLATHEM - PAUG JUNE 2013
3
KEY FIGURES
Mobile Network
Fixed Network
Extensive 2G/3G coverage, with higher
forthcoming H+ speeds (especially 4G
technology)
Bouygues Telecom is currently the only provider of
ADSL, FTTLA (fiber optic with Coax Cable Termination),
and FTTH (fiber to the home)
• 99% of the population has 2G coverage, 96%
3G+ (up to 7.2 Mb/s), and 58% in H+, the
fastest 3G speeds - up to 42 Mb/s
• 4,800 connection nodes
• 15,000 relay antennas
• Almost 7 million households eligible for the very high
fixed speed (THDF) due to an agreement with
Numericable (2009)
• Coverage in more than 250 international
locations due to agreements with more than
530 foreign operators
• Deployment of 4G began in 2012, with
coverage in the city of Lyon
• Over 10 billion euros invested in the network
since 1995
C. VANLATHEM - PAUG JUNE 2013
• 78% of households covered in unbundled zone
• 40,000 km of fiber
• Access to THDF for over 13 million potential
households due to agreements with SFR (2010) and
Orange (2012) in both high and low density zones.
WHY AN APM ON THE FIXED OSS
Supervision based on the
thresholds
IHM Sales/ IHM PNF / IHM Cdc /
Diag / …
Requests
Customer
Reference
BSS
MAJ
Availability of front OSS Sale:
Eligibility / OSS Guichet /
NRM
NAGIOS
information
feedback
OSS Front
PFS Network
(SMV / DSLAM / ACS / IMS
/ Radius HS)
Provisioning
Provisioning
OSS Back
MAJ
Référentiels
Technical
OSS
techniques
standards
OSS
Stream of orders
Availability of the Provisioning
chains
NAGIOS
C. VANLATHEM - PAUG JUNE 2013
Operators/partners Relationship
QOD: flows
Consistency and
OSS ref
5
TOOLS AVAILABLE IN THE APM (APPLICATION
PERFORMANCE MANAGEMENT) SPACE
CRITERIA USED FOLLOWING THE POC.


Criteria for choosing an AMP:

Simple installation and configuration

Available for a variety of technologies (Java, .NET, PHP, etc.)

User-friendly interface for analytics

Cost

Support requirements taken into account
AppDynamics met all of these requirements:

Auto-discovery mechanism

Clear dashboard

Economical model based on the number of JVMs.
C. VANLATHEM - PAUG JUNE 2013
6
APPDYNAMICS AT BOUYGUES TELECOM
 Usage environment

Provisioning FAI

40 JVMs
 Principal uses

Monitoring/Measuring=> Provide visibility

Supervision => Detect incidents more quickly

Diagnostic assistance => Reduce delays in incident resolution

KP assistance / Performance measurement => Anticipate problem points
 Forthcoming tasks to reach our target

Finalize JMX metrics

Create alerts

Interface with NAGIOS

Customize our dashboards

Preview monitoring of our new management system: Fulfillment Order Management Suite
from TIBCO
C. VANLATHEM - PAUG JUNE 2013
7
APPDYNAMICS AT BOUYGUES TELECOM
 Immediate gains with respect to monitoring….
8
APPDYNAMICS AT BOUYGUES TELECOM
 …and with respect to error analysis
9
CONCRETE EXAMPLE FINDING THE CAUSE OF A
BOTTLENECK OF A (ST.)?

Symptom : Bottleneck in the ODYL domain

Functional impact: provisioning optimization process for stalls (?)

Technical mpact : frequent alarms being relayed to the cockpit (support level 1) requiring repeated stopping
and restarting

Process:

Analyzing errors in the Errors dashboard => no significant problem found
C. VANLATHEM - PAUG JUNE 2013
10
CONCRETE EXAMPLE FINDING THE CAUSE OF
A BOTTLENECK OF A (ST.)?


Nothing significant uncovered at the memory level when incidents occurred (Memory dashboard)
Nothing significant uncovered at the response time level when incidents occurred (Slow Response
Times dashboard)
CONCRETE EXAMPLE FINDING THE CAUSE OF A
BOTTLENECK OF A (ST.)?

Observation of different indicators and metrics
available from AppDynamics::

Detection of regular saturation of a connection pool
(?)

Discovery of an anomaly

Correction of the anomaly

Re-establishment of service confirmed by various
measurements
Connection pool utilization on the ST (?)
C. VANLATHEM - PAUG JUNE 2013
Download