1. MOBILE NETWORKS DISASTER RECOVERY USING SDN-NFV
Larysa Globa, Ievgen Volvach
National Technical University of Ukraine “Kyiv Polytechnic Institute”

2. IDEA
Usually mobile network equipment designed to cope with internal disruptions that affect a single element or group of elements of the network infrastructure like a board fault.
There are different methods such as backup, load sharing, creation of resource pool, redundancy configuration, optimized fault detection and isolation techniques for boards and the system.

3. IDEA
But what if external faults created by natural events, such as earthquakes, floods, fires, intentional attacks, etc. happens?
For example, Japan earthquake in 2011 caused cut off millions of users from mobile service and even the emergency restoration took more than one month.
These disruptions are much longer in nature, less probable to happen, and network operators tend to forget them.

4. IDEA
At the same time network operators use a variety of proprietary hardware appliances
(i.e. Serving Gateway (SGW),
Mobility Management Entity (MME),
Packet Data Network Gateway (PGW)
and Home Subscriber Server (HSS)).
Hardware appliances, though deliver performance, are complex to manage, not easy to scale up/down in capacity and not cost effective.

5. REQUIREMENTS So resilient mobile networks must have:
Abundant and redundant network resources for large bandwidth and high reliability.
A part of backbone networks may continue to survive even if a large scale link/node failure occurs.
Utilizing still available network resources (links, nodes) could enable the network to continue providing acceptable services for quick recovery.
Disaster resilience is the capability to absorb the impacts of a disaster, maintain as much function as possible, and make certain that failures occur in a controlled way. This permits as rapid as possible reestablishment of normal or near-normal function.

6. PROPOSED SOLUTION
Schematic example of traffic rerouting in case of disaster

7. PROPOSED SOLUTION
To build effective and disaster-resilient mobile networks, it’s proposed to use SDN (Software Defined Networking) and NFV (Network Function Virtualization) approaches,
which simplifies resource reallocation and minimizes recovery time.

8. PROPOSED SOLUTION
In order to fulfill these idea we propose to add in the network SDN/NFV controller
that is responsible for network monitoring and decision-making
and activating the appropriate scenarios, depending on the nature of the accident.

9. PROPOSED SOLUTION
IF disaster happens the fault registers with automated monitoring system and SDN/NFV controller analyzes the impact of the accident on the service,
and IF the analysis shows the impossibility of restoring the normal operation of the system,
THEN
the scenario is activated and traffic switches to the network element based at the cloud infrastructure.

10. PROPOSED SOLUTION To do this,
SDN/NFV controller compute the amount of necessary cloud resources to run virtual network elements.
The next step is
to rerouting traffic to the newly created element.
After the restore functionality of a network element, that failed, controller reroutes traffic back.

11. PROPOSED SOLUTION
On the first stage of our research we would like to pay more attention to packet data equipment of mobile networks (2G…5G)
After the disaster happens it’s need to control 2 parameters:
Quality of Service (QoS)
Quality of Experience (QoE)

12. PROPOSED SOLUTION To control QoE
It’s necessary to modify control and user traffic flows, so we create new model of service delivery that back up all active subscriber sessions.
So subscribers don’t need to make any actions (like mobile equipment restart or packet data session reinitialization) and after the disaster happens they don’t feel any problems with service.

13. PROPOSED SOLUTION To control QoS
As network has bandwidth limitations, after traffic rerouting it’s possible that user traffic is greater than the traffic threshold supported by system, for that case we propose to use integrated parameter KQI (Fig. on next slide)
In case of KQI downgrade we propose to use traffic prioritization technique to block “heavy” services such as video streaming and mark services like messengers with high priority

14. Parameters used for KQI computing

15. Additional controlled parameters
network survivability;
network availability;
network reliability;
he probability of packet loss;
jitter;
software failure;
security ……….and others