A mechanism of security monitoring and management for service resources in Computing-Aware Traffic Steering (CATS)

Introduction As described in [I-D.ietf-cats-usecases-requirements], traffic steering that takes into account both the computing resource metric and network metric would improve the QoE of several services, e.g., AR/VR and intelligent transportation. But when executing services, efficiency is not the only factor to consider. Security requirements are important for users, service providers and network providers, such as following security requirements: Many resolutions need to be designed and applyed to fully meet security requirements above. But a basic resolution is to monitor and discover corrupted service resources and limit them from providing services and interact with network components as well as other service resources, and more secure service resources should be selected under same conditions. This draft proposes a mechanism to realize monitoring and management of service resources.

Problem statement Service resources face kinds of attacks from inside and outside. Successful attacks may bring security risks for services, such as service unavailable, service data stolen etc, and service resources may become attack tools to interfere activities of other computing resources and functional components in CATS if they are manipulated maliciously. The more vulnerable the resource, the greater the likelihood of being breached. And when service resources have been breached, they will be threats in network. For protecting service availability, security of services and maintaining network stability, computing resources in CATS should be in security status and services should run on reliable resources. So there should be a mechanism to monitor security status of service resources and discover the resources that are likely to be breached and those that have already been breached in CATS. Then different policies can be applyed for above service resources such as excluding the service resources from CATS or reduce usage of them.

Framework and Components In [I-D.ldbc-cats-framework], a CATS framework for computing-aware traffic steering according to computing metrics of service resources and network metrics is proposed. In this draft, the new functional component C-SeMA is introduced based on the existing CATS framework. The CATS Security Metric Agent(C-SeMA) is responsible for collecting security information of computing resources that used to carry service instances. Computing resources could be servers, virtual machines or containers. Details of security information are listed in 5.2. The C-SeMA gathers the security information and decides the security status of the computing resources, then generates security policy according to the security status of computing resources and sends it to C-PS. C-PS could adjust resource selection policy according to the security policy.

Workflow

Overview Figure 2 shows the main workflow of monitoring and management of computing resources. Two stages are included in the workflow. In stage 1, C-SeMA acquire security information from computing resources or security functions. In stage 2, security policies will be generated according to security information and will be sent to C-PS. In stage 3, C-PS can select service nodes according to the security policies. | |security | | |information | | | | | | processing of | | security | | information | | ++---------------------------+ | | security policy | | | | | | service | | decision-making | | | Figure1: overview of workflow ]]>

Collection of security information C-SeMA need to collect security capability information and security status information of computing resources. Security capability information indicates the security capabilities that are depolyed for protecting computing resources and services running on them from attacks. Security status information including predefined dynamic security metrics of computing resources, such as abnormal traffic, adnormal behaviors, and vulnerability situation etc. C-SeMA need to distribute requirements of security metrics to computing resources or security functions. C-SeMA collects security information from computing resources or security functions deployed in service sites or network. Different collection methods can be used, such as C-SeMA collects security information periodically or security function push the information when security status changes. Security information should be associated with a specific computing resource and is associated with service instances by computing resources. Table 1 shows the examples of security information for service resources.

Processing of security information C-SeMA sets classification rules for computing resources. The classification rules classify computing resources into different security priority according to security information, such as low, middle, high and disabled. C-SeMA generates initial priority of computing resources when receiving security information for the first time and adjust the priority when security information update. Two examples of classify rules are as following: Examples of Initial priority of computing resources are as follows: C-SeMA processes the updated security information and determines the priority adjustment policy to allocate new priority for computing resources or service instances. The policies include but not limited to the following types: Specific decision logic is out of the scope of this draft.

Service decision-making based on security information C-SeMA submits initial security priority and updated priority policies for specific computing resources or service instances to C-PS. C-PS could take security priority of computing resources as references and adjust selection policies when conducting service decision-making:

IANA Considerations This memo includes no request to IANA.

Security Considerations TBD