Network Working Group | Z. Li |
Internet-Draft | S. Peng |
Intended status: Standards Track | Huawei Technologies |
Expires: September 7, 2020 | D. Voyer |
Bell Canada | |
C. Li | |
China Telecom | |
L. Geng | |
China Mobile | |
C. Cao | |
China Unicom | |
K. Ebisawa | |
Toyota Motor Corporation | |
S. Previdi | |
Individual | |
J. Guichard | |
Futurewei Technologies Ltd. | |
March 6, 2020 |
Application-aware Networking (APN) Framework
draft-li-apn-framework-00
A multitude of applications are carried over the network, which have varying needs for network bandwidth, latency, jitter, and packet loss, etc. Some new emerging applications (e.g. 5G) have very demanding performance requirements. However, in current networks, the network and applications are decoupled, that is, the network is not aware of the applications' requirements in a fine granularity. Therefore, it is difficult to provide truly fine-granularity traffic operations for the applications and guarantee their SLA requirements.
This document proposes a new framework, named Application-aware Networking (APN), where application characteristic information such as application identification and its network performance requirements is carried in the packet encapsulation in order to facilitate service provisioning, perform application-level traffic steering and network resource adjustment.
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."
This Internet-Draft will expire on September 7, 2020.
Copyright (c) 2020 IETF Trust and the persons identified as the document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
A multitude of applications are carried over the network, which have varying needs for network bandwidth, latency, jitter, and packet loss, etc. Some applications such as online gaming and live video streaming has very demanding network requirements and therefore require special treatment in the network. However, in current networks, the network and applications are decoupled, that is, the network is not aware of the applications' requirements in a fine granularity. Therefore, it is difficult to provide truly fine-granularity traffic operations for the applications and guarantee their SLA requirements accordingly. [I-D.li-apn6-problem-statement-usecases] describes the challenges of traditional differentiated service provisioning methods, such as five tuples used for ACL/PBR causing coarse granularity, DPI imposing high CAPEX & OPEX and security issues, as well as orchestration and SDN-based solution causing long control loops.
This document proposes a new framework, named Application-aware Networking (APN), aiming to guarantee fine-granularity SLA requirements of applications, where application characteristic information such as application identification and its network performance requirements is carried in the packet encapsulation in order to determine the path, steer traffic, and perform network resource adjustment.
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119.
This document is not a protocol specification and the key words in this document are used for clarity and emphasis of requirements language.
ACL: Access Control List
APN: Application-aware Networking
APN6: Application-aware Networking for IPv6/SRv6
DPI: Deep Packet Inspection
MPLS: Multiprotocol Label Switching
PBR: Policy Based Routing
QoE: Quality of Experience
SDN: Software Defined Networking
SLA: Service Level Agreement
SR: Segment Routing
SR-MPLS: Segment Routing over MPLS dataplane
SRv6: Segment Routing over IPv6 dataplane
The APN framework is shown in Figure 1. The key components include Service-aware App, App-aware Edge Device, App-aware-process Head-End, App-aware-process Mid-Point, and App-aware-process End-Point.
Packets carry application characteristic information (i.e. application-aware information) which includes the following information:
Client Server +-----+ +-----+ |App x|-\ /->|App x| +-----+ | +-----+ +---------+ +---------+ +---------+ | +-----+ \->|App- | |App-aware|-A-|App-aware|-A-|App-aware|-/ User side |aware|-|process |-B-|process |-B-|process | /->|Edge | |Head-End |-C-|Mid-Point|-C-|End-Point|-\ +-----+ | +-----+ +---------+ +---------+ +---------+ | +-----+ |App y|-/ \->|App y| +-----+ --------- Uplink ----------> +-----+
Figure 1: Framework and Key Components
The key components are introduced as follows.
In this way the network is aware of the service requirements expressed by the applications explicitly. According to the service requirement information carried in the packets the network is able to adjust its resources fast in order to satisfy the service requirement of applications. The flow-driven method also reduces the challenges of interoperability and long control loop.
APN doesn't mandate a specific encapsulation however it is reasonable to assume that most of the APN benefits are achieved when utilizing IPv6 encapsulation (e.g. IPv6 header as well as, possibly, extension headers). APN6 (the APN architecture applied to the IPv6/SRv6 data plane) consists of the application-aware information conveyed into the network through the use of IPv6 header and Extension Headers and where the network performs service provisioning, traffic steering, and SLA guarantee according to such information. This section specifies the requirements for supporting the APN framework, including the requirements for conveying and handling the application-aware information and related security requirements. Other encapsulation may be used with some obvious constraint such as, as in the case of MPLS, the limited space available in the header (i.e., 20-bit label size).
The application-aware information includes application-aware identification information and network performance requirements information.
The different combinations of the IDs can be used to provide different granularity of the service provisioning and SLA guarantee for the traffic.
The different combinations of the parameters are for further expressing the more detailed service requirements of an application, conveyed together with the Application-aware identifiers, which can be used to match to appropriate tunnels/SR Policies, queues that can satisfy these service requirements. If not available, new tunnels/SR Policies can also be triggered to be set up.
[REQ 1a]. Application-aware identification information MUST include Application ID to indicate the application that generates the packet.
[REQ 1b]. SLA level is RECOMMENDED to be included in the Application-aware identification information.
[REQ 1c]. User ID and Flow ID are OPTIONAL to be included in the Application-aware identification information.
[REQ 1d]. Network performance requirements information is OPTIONAL.
[REQ 1e]. All the nodes along the path SHOULD be able to process the application-aware information if needed.
[REQ 1f]. The application-aware information can be generated directly by application, or by the application-aware edge devices though packet inspection or local policy.
[REQ 1g]. The application-aware information SHOULD be kept intact when directly copied from the application-aware edge devices and carried in the packet.
The app-aware-process Head-End and app-aware-process Mid-Point perform matching operation against the application-aware information, that is, to match IDs and/or service requirements to the corresponding network resources (tunnels/SR policies, queues).
In order to achieve better Quality of Experience (QoE) of end users and engage customers, the network needs to be able to provide fine-granularity and even application-level SLA guarantee [I-D.li-apn6-problem-statement-usecases].
[REQ 2-1a]. With the application-aware information, the App-aware-process Head-End SHOULD be able to steer the traffic to the tunnel/SR policy that satisfies the matching operation.
[REQ 2-1b]. With the application-aware information, the App-aware-process Head-End SHOULD be able to trigger the setup of the tunnel/SR policy that satisfies the matching operation.
[REQ 2-1c]. With the application-aware information, the App-aware-process Head-End and Mid-Point SHOULD be able to steer the traffic to the queue that satisfies the matching operation.
[REQ 2-1d]. With the application-aware information, the App-aware-process Head-End and Mid-Point SHOULD be able to trigger the configuration of the queue that satisfies the matching operation.
Network slicing provides ways to partition the network infrastructure in either control plane or data plane into multiple network slices that are running in parallel. The resources on each node need to be associated to corresponding slices.
[REQ 2-2a]. With the application-aware information, the App-aware-process Head-End SHOULD be able to steer the traffic to the slice that satisfies the matching operation.
[REQ 2-2a]. With the application-aware information, the App-aware-process Mid-Point SHOULD be able to associate the traffic to the resources in the slice that satisfies the matching operation.
Along the path each node needs to provide guaranteed bandwidth, bounded latency, and other properties relevant to the transport of time-sensitive data for the Detnet flows that coexist with the best-effort traffic.
[REQ 2-3a]. With the application-aware information, the App-aware-process Head-End SHOULD be able to steer the traffic to the appropriate path that satisfies the matching operation.
[REQ 2-3b]. With the application-aware information, the App-aware-process Head-End SHOULD be able to trigger the setup of the appropriate path that satisfies the matching operation for the Detnet flows.
[REQ 2-3c]. With the application-aware information, the App-aware-process Mid-Point SHOULD be able to associate the traffic to the resources along the path that satisfies the performance guarantee.
[REQ 2-3d]. With the application-aware information, the App-aware-process Mid-Point SHOULD be able to reserve the resources for the Detnet flows along the path that satisfies the performance guarantee.
The end-to-end service delivery often needs to go through various service functions, including traditional network service functions such as firewalls, DPI as well as new application-specific functions, both physical and virtual. SFC is applicable to both fixed and mobile networks as well as data center networks.
[REQ 2-4a]. With the application-aware information, the App-aware-process devices SHOULD be able to steer the traffic to the appropriate service function.
[REQ 2-4b]. The App-aware-process devices SHOULD be able to process the application-aware information carried in the packets.
Network measurement can be used for locating silent failure and predicting QoE satisfaction, which enables real-time SLA awareness/proactive OAM.
[REQ 2-5a]. With the application-aware identification information, the App-aware-process devices SHOULD be able to perform IOAM based on the Application ID.
[REQ 2-5a]. With the application-aware information, the network measurement results can be reported based on the Application ID and verify whether the performance requirements of the application are satisfied.
[REQ 3a]. The security mechanism defined for APN MUST allow an operator to prevent applications sending arbitrary application-aware information without agreement with the operator.
[REQ 3b]. The security mechanism defined for APN MUST prevent an application requesting a service which it is not entitled to get.
This document does not include an IANA request.
[I-D.li-apn6-problem-statement-usecases] and describe the security considerations and requirements for APN.
The authors would like to acknowledge Robert Raszuk (Bloomberg LP) and Yukito Ueno (NTT Communications Corporation) for their valuable reviews and comments.
Daniel Bernier Bell Canada Canada
Email: daniel.bernier@bell.ca
Chongfeng Xie China Telecom China
Email: xiechf@chinatelecom.cn
Peng Liu China Mobile China
Email: liupengyjy@chinamobile.com
Zhuangzhuang Qin China Unicom China
Email: qinzhuangzhuang@chinaunicom.cn
Chang Liu China Unicom China
Email: liuc131@chinaunicom.cn
[I-D.li-apn6-problem-statement-usecases] | Li, Z., Peng, S., Voyer, D., Xie, C., Liu, P., Liu, C., Ebisawa, K., Previdi, S. and J. Guichard, "Problem Statement and Use Cases of Application-aware IPv6 Networking (APN6)", Internet-Draft draft-li-apn6-problem-statement-usecases-01, November 2019. |
[RFC2119] | Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/RFC2119, March 1997. |
[RFC7665] | Halpern, J. and C. Pignataro, "Service Function Chaining (SFC) Architecture", RFC 7665, DOI 10.17487/RFC7665, October 2015. |
[RFC8200] | Deering, S. and R. Hinden, "Internet Protocol, Version 6 (IPv6) Specification", STD 86, RFC 8200, DOI 10.17487/RFC8200, July 2017. |
[RFC8578] | Grossman, E., "Deterministic Networking Use Cases", RFC 8578, DOI 10.17487/RFC8578, May 2019. |
[RFC3272] | Awduche, D., Chiu, A., Elwalid, A., Widjaja, I. and X. Xiao, "Overview and Principles of Internet Traffic Engineering", RFC 3272, DOI 10.17487/RFC3272, May 2002. |