Cloud Native Computing Foundation Cloud Native Telco Day EU 2022, 19 May 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Lightning Talk: 5G CNFs Observability Based on eBPF - Abderaouf Khichane & Ilhem Fajjari, Orange

Description

Lightning Talk: 5G CNFs Observability Based on eBPF - Abderaouf Khichane & Ilhem Fajjari, Orange

Monitoring the traffic between the microservices of a 5G network service is of great interest, as the network service can only provide its functionality in most cases if several microservices are involved. In addition to ensuring that the state of the exchanges is good, tracking the exchanges between these microservices allows to retrieve global information on the state of the system. Several solutions in the open source ecosystem allow to track exchanges in microservices architectures. These solutions are generally based on service mesh technology and are more focused on handling exchanges based on protocols widely used in the IT world such as HTTP/1.1, HTTP/2 and gRPC. The use of such a model based on the injection of proxies that will coexist with the network service components will generate an overhead in the latency of the exchanges. The question is how to monitor the exchanges between the components of a 5G network service that are based on protocols widely used in IT such as HTTP/2 but also telco specific protocols such as NGAP, PFCP and NGAP without changing their source code and with a minimum overhead?

A

So thank you for uh being present today for uh for our presentation on 5g cnf observability based on ebpf, so I will start by presenting myself I'm abdarov kishan, I'm a phd student in orange labs and paris sacre university.

A

I specified that this presentation was prepared jointly with elon fajri, who is a research project leader on cloud native network functions orchestration in orange labs. Unfortunately, she she cannot be present today due to health conditions.

A

Okay, so for the plan of this presentation, I will start by uh presenting the challenges and the objective of our work. Then I will present our ebpf based solution for the observability of 5g cloud native network functions.

A

Finally, I will describe the x the experimentations. We did with our solution before concluding the presentation.

A

Okay, so our work is, uh is mainly related to 5g and the observability of its network functions so, as you see for the 5g architecture is highly distributed.

A

So that means that, in order to provide an end-to-end functionality, multiple network functions will be involved, and so it is useful to uh to uh it is. It is important to monitor the traffic. The network traffic between these network functions first to uh to uh to to ensure a good level of uh performance by talking the, for example, the error rate and request processing, but also to get informations about the the global state of the 5g system like, for example, the number of of attached users or, and the number of created video sessions.

A

So in order to provide functionalities, 5g network functions communicate between with each other using a variety of protocols. So some of them are originating from I.t like http, for example, while others are typical. Specific and artists are are defined by 3gpp specifications like ngap and pfcp for senioric messages and gtpu to forward the end users, traffic.

A

Okay, so around this context and challenges, the the main objective of our work is the is to implement a framework for the observability of network traffic between 5g cloud native network functions or 5g cnfs.

A

The id the idea behind that is to sniff the traffic perform protocol analysis and expose metrics.

A

While designing our solution, we are considering the opera, the open source ecosystem, so our solution is designed to be deployed on top of the kubernetes platform, which is uh which is uh used for the orchestration of cloud native network functions in general.

A

Besides, we are leveraging prometheus and grafana for collecting for collecting, storing and visualizing metallics.

A

Finally, we are using ebpf, so ebpf is uh is a technology which is originated from from the linux kernel and it allows to run event driven programs based on space on specific events,.

A

Okay, so now, let's, uh let's give an overview of our solution. So, as already mentioned, our solution works on top of kubernetes platform and allows to monitor the traffic between kubernetes ports that are, that belong to a 5g network service. To do so, we def we developed two components.

A

First, the controller is uh is, uh is managed by equipment, is deployment and is responsible of discovering cuban sports by communicating with the the ipi server. Secondly, the filter is is managed by a diamond set cabinet is demonstrated, and so it is present on each node within the capacities cluster.

A

I will give more details on its role, but all you have to know, for instance, is that it is responsible of monitoring the traffic between the discovered ports and exposing metallics.

A

These metrics can be then scrapped by by parameters and display displayed on the graph on graphene dashboards.

A

Okay, so now, let's uh focus on uh on our filter, so we use uh we used python to develop it, and we are also using the vcc frame framework which is providing a library allowing to manage ebpf programs.

A

So I will call this our user space program. It will based on the information provided by the controller. It will attach ebpf programs to network interfaces on the worker node and which are linked to network interfaces on the pod. So I specify that also we can. We can monitor the traffic.

A

We can monitor uh ports with multiple network interfaces.

A

These bpf programs will continuously filter, 5g network events and send them again to the user space program.

A

The letter will perform protocol analysis to ex to extract information depending on on the protocol, and then we display logs and expose metallics so to test our solution.

A

We we made use of 5g open source projects, including 45 dc and open 5gs for the core network part, and we are also using ul and sim to which is the project providing a simulation of 5g, node, b and ue to uh to trigger to trigger the exchange of network messages.

A

In this slide, we can see examples of metrics that we are exposing with our solution. So the first one is our in general, uh the metrics for for request and response based protocols. We are tracking the the total number of requests, the total number of responses and the response time.

A

These methods are labeled by uh the server and client context. So a context include like the the name of the pod, uh the name of the interface, the ip address, the name of the deployment the port belong to and all the liables which are related to the protocol itself, so like, for example, for pfcp the the message, the type of the message and also the close, which represents the state of the response.

A

So it can be a success or the cause of errors of the error in case of errors for other protocols like gtpu, we are tracking the total number of messages, the total number of bytes, but also the forwarding time so for the upf, for example, it acts like a router which forward the messages from an interface to another to another one, and so we are tracking this uh this time.

A

To conclude this presentation, so in this talk with, uh I described our solution for uh for for the observability of 5gcnf based on ebpf.

A

There is an in progress work to uh to test our solution with the large amount of data by triggering the the simulation of a large number of user equipments and also to handle encrypted traffic.

A

Mainly the protocol https later we we aim to to test our solution with the with the data acceleration technologies for the upf, like, for example, when sruv is used for the upf and also to leverage meet the exposed metrics and logs using artificial intelligence.

A

Okay. So, by the end of this presentation, I would like to thank you for your attention and I'm available for any question or remark.