Você está na página 1de 8

Guaranteeing Service Assurance in NFV

WHITE PAPER
DECEMBER 2016
Guaranteeing Service Assurance in NFV

NFV comes with great promises

Legacy networks are notoriously difficult to modify or upgrade since physical hardware is
expensive, proprietary and inflexible. Network Functions Virtualization (NFV) proposes
a fundamental new way to define, create and manage a network by replacing dedicated
network hardware with software & automation.

There is no doubt about CSPs moving forward with NFV but its likely to take a decade,
and likely to happen in steps rather than as a rip-and-replace of the infrastructure. IHS
Markit forecasts that the NFV market will be worth $15.5 billion by 2020 and that the NFV
market will grow at a compound annual growth rate of 42 percent between 2015 and
2020 from $2.7 billion in 2015 to 15.5 billion in 2020.

NFVs value proposition is overwhelming:

Increased network flexibility, scalability & reliability

Direct CAPEX savings through the use of COTS hardware:


1. Reduction in hardware, rack space and power consumption
2. Reuse of a shared pool of hardware resources for various
functions to meet elastic capacity needs

Automated provisioning of network applications and associated


sustainable OPEX reductions

Improved and on-demand service deployment

According to an Analysys Masons report of February 2015, mobile operators can for
instance achieve 33% net savings by implementing a vIMS for VoLTE, 45% in net
capex savings and 10 in net opex savings.

33% vIMS net cost savings

45% 10%
net less capex net less opex

25% 28%
75% from
25% 17%
faster and from 30%
from impro- automated
from cheaper from
ved service service from data
improved VNF crea- reduced
assurance fulfillment centre cost
capacity tion and customer
T2R and inventory savings
planning upgrades complaints
O&M and optimi-
on COTS
VNI sation

-2-
but also with big challenges

NFV is a disruptive technology for network design and service delivery. It introduces
new risks because it adds additional layers of complexity, which need to be tracked and
managed.

Increased complexity

The introduction of a virtual infrastructure with an orchestrator that automates the in-
troduction of new network functions brings a real complexity to network management.
CSPs need to understand which of the decisions triggers automation and how to mea-
sure their impact.
Additionally, these virtual network functions must integrate perfectly well with physical
network functions. So, new solutions will be necessary to help CSPs manage service
assurance in this context of hybrid networks.

Organizational transformation

NFV requires an organizational transformation since it implies a shift towards IT. Network
engineers and support teams that are used to working with physical infrastructure will
need to adapt their skills and competences, redesign existing processes and methods
and manage new tools to enable the organization to migrate effectively and seamlessly
to virtualization.

Impact on the customer experience


Within NFV, new services and capacity requirements can be changed on the fly to meet
demand, but are quite risky to subscribers QoE if managed incorrectly. To avoid the risk
of service disruptions and bad customer experience when moving to a software-based
network architecture, providing a good service assurance should not be overlooked.

NFV standardization provides already the integration of automatic active end-to-end


testing in the workflows of new services implementation. However, this type of automa-
tion only provides a minimum guarantee that the network service is properly instantiated
and works within the predefined configurations. It does not measure the impact on the
customer experience. Having a good way to test and validate is therefore paramount.

-3-
Guaranteeing Service Assurance in NFV

Service Assurance is business critical

Maintaining subscriber experience is a major challenge especially with the rapid growth
of rich media applications and exploding network traffic. According to Analysys Mason
churn in mature markets is as high as 24%, so there is a real need to react to customer
issues in real-time, or even better, to proactively move on the issue before customer
realizes there is one. Service assurance is an area of great importance and should be
an integral part of service delivery, in a physical and virtualized environment.

Relevance of network probes

Monitoring and optimizing performance in a virtualized environment is more challenging.


It complicates the implementation of probes since the monitored interfaces can potenti-
ally be virtual (East-West interfaces).

Instantiation of vProbes
A second difficulty is linked to the network orchestrator which can, under certain condi-
tions, decide to modify network configuration by moving, deleting or adding new network
nodes and hence add or remove virtual links. To cope with this network elasticity, probes
are going virtual as well. Monitoring solutions hence become themselves a VNF as part
of the virtual network. In this case probes are instantiated by the orchestrator at the
same time as network creation. Though, at this moment, standards which describe how
to duplicate network flows towards monitoring solutions when creating of a new network
service arent availble yet.

Flow duplication
Astellia therefore proposes the integration of a TAP Manager application in its mo-
nitoring solution which detects any change in the network and instructs the network
controller to duplicate the relevant flows to virtual probes. Astellias Network Topology
Manager can detect if a new VM instance has been launched for a new VNF and can
then reconfigure its monitoring solution to instantiate a new vProbe and reconfigure the
tapping to monitor this new VNF. Another component Astellia offers is a VNF Manager
that takes care of the health of the monitoring solution and is in charge of automatically
starting a new vProbe in case of failure or auto-scaling the system by deploying additio-
nal vProbes when processing requirements increase.

-4-
Guaranteeing Service Assurance in NFV

Figure 1: Provisioning vProbes

Orchestrator 6 Analytics

VNF Manager VNF Manager TAP


VNF 1 VNF 2 vProbe
Equipment Monitoring Manager

1 2 3 4

Network Virtual Infrastructure


Virtual Infrastructure Manager
Manager Network Infrastructure

1. The orchestrator books resources at the Virtual Infrastructure Manager


2. The orchestrator asks the VNF manager to provision VNFs
3. The orchestrator asks to provision a monitoring solution on the same pop
4. The orchestrator provisions network resources for interconnecting VNFs
5. The orchestrator notifies the TAP Manager that the network topology has
changed
- the TAP manager collects topology information
- the TAP manager provisions flow duplication rules to forward and tun
nels traffic to be monitored to the vProbe
6. Decisions based on analytics are fed into the orchestrator

-5-
Guaranteeing Service Assurance in NFV

Key role of advanced analytics

The innovations brought by NFV offer also new ways of monitoring, focusing for example
on key interfaces to quickly identify faulty network equipment. Software probes can then
be instantiated on demand to ensure a more precise diagnosis in order to resolve the
most complex problems.
Its inevitable that NFV will create a challenging environment for maintaining a consistent
QoE. Network analytics will therefore play a key role in managing service assurance in
NFV networks. They provide tangible insights and enable proactive decision-making
needed to improve customers experience, reduce network downtime, optimize network
planning and reduce costs.

Analytics are critical to offer for instance real-time intelligence to dynamically scale re-
sources to cope with unexpected changes in network traffic. Analytics solutions should
therefore combine different sources of information: vProbes, OSS counters, active tes-
ting, NetFlow, virtual infrastructure counters and orchestration orders.

Monitoring E2E Services Service quality


demarcation
Alarms

Actions

Root cause analysis

Correction
Troubleshooting recommendation

Analysis & Analytics

Collection layer

Active Virtual Orches-


OSS vProbes NetFlow
testing infra tration

In the near future, analytics will directly feed actionable intelligence decisions to the
orchestrator by combining information on the functioning of VNF, service performance
and customer experience. Analytics can bring this subscriber insight to the orchestrator
and trigger subscriber-centric actions, for instance propose rules to better serve VIP
subscribers in congested areas.

-6-
Astellias 3 must-follow guidelines for
implementing Service Assurance in NFV

1. Monitor the network in real-time with vProbes

Not all network functions will be virtualized and migration to a virtual infrastructure will
be a gradual process. Maintaining an end-to-end visibility, both within virtual infrastruc-
ture as well as across hybrid physical and virtual environments, is vital to guarantee
subscribers do not experience anything less than the best performance on the network.
Astellia recommends the use of vProbes to monitor physical and VM-to-VM communi-
cations and ensure a real-time view of the entire network environment to predict network
problems and put processes in place to ensure that high value customers consistently
receive a good QoE.
It is crucial to monitor and compare QoE trends before and after the migration towards
NFV. Astellia guarantees a seamless migration by depicting the same KPIs in the same
graphical user interface to quantify gains or gaps in QoE.

2. Implement Service Assurance based on analytics

E2E issues demarcation, fault isolation and root cause analysis make service quality
management more efficient. Issue demarcation is an automatic diagnosis based on ma-
chine learning algorythms which provides a list of probable failure causes.
Astellias analytic solution, backed-up by a holistic and proven portfolio of professional
services, integrates key metrics from network infrastructures and VNF, allowing for a
complete consolidated vision and control of virtualized networks.

3. Partner with the right solution provider

For the last two years, Astellia has been working on its solution to address current and
future challenges of NFV. Its new vProbes, compatible with the most important stan-
dards of the cloud, allow easy implementation and integration in a virtual platform. In
its short term roadmap, Astellia proposes an integration of its entire solution in a fully
orchestrated network.

There is industry consensus that service assurance cannot be an after-


thought when it comes to the operationalisation of NFV/SDN based virtual
networks. As more parts of the network become virtualised, CSPs will need
new assurance approaches and innovative technologies such as vProbes
to monitor and assure services in the NFV domain. Equally important will be
the ability of such a solution to dynamically self-adapt and reconfigure to the
changing network conditions. In this market context, Astellias develop-


ment of the orchestrated virtual probes is important, and strongly posi-
tions the company as an innovator in service assurance for NFV/SDN.
Anil RAO, Senior Analyst at Analysys Mason

-7-
Astellias Service Assurance solution for hybrid networks

BUSINESS INTELLIGENCE MONITORING & TROUBLESHOOTING OPTIMIZATION

Network - Service - Customer

DATA LAKE

DATA ACQUISITION: vProbes, probes, call traces, CRM, billing, etc.

HYBRID MOBILE NETWORK: RAN + CORE

Astellias real-time Nova monitoring and troubleshooting solution for multi-vendor hybrid mobile networks brings unique business intelligence to
reduce OPEX and CAPEX, optimize end-to-end network performance, service quality and enhance subscriber satisfaction. Nova meets requirements
of network operations, service operations, customer care and marketing teams. Nova provides operators with capabilities to detect, correlate, analyze,
report and troubleshoot issues related to network performance, handset behavior and subscriber usage.

Astellias vProbe
Astellias vProbe delivers the same high value information as its physical probes, using the same core software and DPI technology while being adap-
ted to a virtual environment through the integration of a new capture layer based on DPDK libraries and high precision soft timestamping functions.
Astellias dimensioning model allows to calculate very precisely the resources needed for the vProbe (CPU, RAM, storage IOPS, storage capacity and
network bandwidth) based on throughput requirements and details of the deployment site.

The table below illustrates the typical resources required by a vProbe for Core LTE interface monitoring.
Captured interfaces: S1-C, S6a, S11, S10, S5/S8, Gn-Core, Gn-Mobility
vProbe (Neptune)
Input deployment RAM
# Vcpu@ 2.6 GHz Storage size (TB) Storage IOPS Output BW (MB/s)
requirements (GB)
Minimal configuration:
100 Mbps Control
4 40 10 60 10
Plane
(7 days of history)
The dimensioning model is automated in an Astellia management module that resizes the monitoring solution upon the topology of customer deploy-
ment.

About Astellia
Astellia is a leading provider of network and subscriber intelligence enabling mobile operators to drive service quality, maximize operational efficiency,
reduce churn and develop revenues. Its vendor-independent real-time monitoring and troubleshooting solution optimizes networks end-to-end, from
radio to core. Astellias unique blend of products and services provides automated optimization, actionable geolocated insights and big-data analytics
to Network Operations, Service Operation Center, Customer Care and Marketing teams. Astellia has close partnerships with more than 120 telecom
operators. Headquartered in France, Astellia is based in Canada, Lebanon, Mexico, Morocco, Russia, South Africa, Spain and the USA.
Follow Astellia: astellia.com, LinkedIn and @Astellia_news.

ZAC Airlande | 2 Rue Jacqueline Auriol BRAZIL | CANADA | LEBANON | MOROCCO


CS 69 123 | 35091 Rennes Cedex 9 | France RUSSIA | SOUTH AFRICA | SPAIN | USA
Tel: +33 299 048 060 | Fax: +33 299 048 061
infos@astellia.com | www.astellia.com

Você também pode gostar