| Internet-Draft | Hybrid Solution for CATS | October 2023 |
| Yi, et al. | Expires 25 April 2024 | [Page] |
Computing-Aware Traffic Steering (CATS) is a traffic engineering architecture that takes the dynamic changes of computing and network resources into account when forwarding traffic to appropriate service instances for processing. For the development of the current network, it is important to have a solution that meets different types of service requirements and can be deployed reasonably. Therefore, this document proposes a hybrid solution to provide differentiated and flexible traffic streering capabilities for different service while saving the cost of retrofitting existing network equipment.¶
This note is to be removed before publishing as an RFC.¶
Discussion of this document takes place on the Computing-Aware Traffic Steering Working Group mailing list (cats@ietf.org), which is archived at https://mailarchive.ietf.org/arch/browse/cats/.¶
Source for this draft and an issue tracker can be found at https://github.com/VMatrix1900/draft-yi-cats-hybrid-solution.¶
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 25 April 2024.¶
Copyright (c) 2023 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 Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
CATS enables large-scale interconnected collaboration at the edge, providing optimal service access and load balancing to adapt to dynamic service. The computing power and network based on the actual processing delay condition can dynamically process the service request to switch to the appropriate service node, thereby improving the quality of service resource utilization and user experience[I-D.ldbc-cats-framework]. CATS needs to provide diversified and differentiated service capabilities so that different service traffic can be forwarded to appropriate service instance. This document proposes a hybrid solution, on the one hand, it uses centralized computing information awareness and distributed routing decision. On the other hand, it can provide the service capability of distributed routing or centralized routing for different services to achieve service differentiation.¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
There are three main aspects of CATS 'work: - Computing information awareness: The network of the resource utilization status and service status of service instanct needs to be notified, so that the network can perceive the status of service contact instance. - Select the optimal service contact instance : The optimal service instance needs to be calculated based on the status of computing and network. - Calculate the optimal forwarding path: After determining the optimal service instance, the optimal forwarding path to the service needs to be calculated. The optimal forwarding path can be determined based on network factors such as delay, packet loss rate, and bandwidth.¶
In the implementation of the above work, CATS still has some problems: - A number of devices will be upgraded and the cost will be high, if the computing information needs to be notified between service instance and engress router. - As business scenarios become more and more diverse, CATS needs to provide differentiated network and computing capabilities for different requirements of different businesses.¶
This document designs a hybrid solution from two aspects: - Collect the computing information through the cloud management platform, and then process and send it to the network device on demand. This kind of centralized computing information awareness is more easier to achieve and less costly to deploy compared to collect the computing information distributedly. - Distributed or centralized routing decision-making methods for different services. For intelligence transportation：The location of vehicles is constantly changing when vehicles are driving normally on the road. CATS needs to re-select the optimal service instance and the optimal forwarding path according to the latest vehicle location. In addition, the scenario of intelligence transportation has very high requirements on delay, and the delay value will directly affect the driving safety of vehicles. Therefore, it is recommended that the ingress router of CATS makes routing decision, rather than the centralized controller recalculating and sending to the ingress router. In this way, the signaling transmission time between the centralized controller and the ingress router is reduced. Thus, the service quality is improved. For other scenario such as VR/AR, SDWAN, which generally prioritize global utilization, it is suitable to use the centralized routing decision method since the centralized controlller has a global perspective.¶
+--------------+ +------------------+
| network | | cloud management |
| controller |<------| platform |
+--------------+ +------------------+
/ \
+------------------+ +---------------+
| R2 R3 |------|service instance|
| | +----------------+
Client---|R1(ingress router)|
| | +---------------+
| R4 R5 |------|service instance|
+------------------+ +---------------+
During the deployment of the service, the cloud management platform or other upper-layer systems collect computing information, process it, and then send it to the required network devices, which are generally network ingress router. Then, CATS determines whether to use the centralized routing decision-making mode or the distributed routing decision-making mode based on the service type. Generally, it is recommended that services with strict requirements on time delay adopt the distributed routing decision making mode, and the other services adopt the centralized routing decision making mode.¶
When the distributed routing decision-making mode is adopted, the CATS ingress router selects the optimal service instance and calculates the optimal forwarding path according to the status of computing and network, and then directs the user traffic to the path towards the service instance.¶
When the centralized routing decision-making mode is adopted, the network controller selects the optimal service instance and calculates the optimal forwarding path according to the status of computing and network, and sends the results to the ingress router. The ingress router directs user traffic to the path towards the service instance.¶
The service instance reports the computing information to the cloud management platform.¶
The cloud management platform processes computing information and send it to the network controller.¶
The network controller collect the network information and process it along with computing information.¶
Distributed routing decision mode: The network controller sends computing information to the network ingress router. The ingress router selects the optimal service instance and calculates the optimal forwarding path.¶
Centralized routing decision mode: The network controller selects the optimal service instance, calculates the optimal forwarding path and sends the result to the ingress router.¶
The ingress router performs traffic steering.¶