Multi-cloud Security - Issue #2314

As cloud infrastructures become a key element in the telecommunication operator ecosystem, particularly with the 5G rollout becoming a critical business necessity, cloud-focused threats and associated adversarial behaviours, methods, tools, and strategies that cyber threat actors use to plan and execute are part of the attack surface landscape.

In an ecosystem comprised of different network domains, products and business partnerships, the responsibility for managing these different cloud environments necessary to support 5G use cases falls to different organisations, creating a new level of complexities and a new range of security risks. In such an environment, there are additional security principles to be considered.  These principles, see Table 2 below, drawn from the collaboration with GSMA FASG [E], intend to address typical cloud associated attacker behaviours, as identified in the widely accepted MITRE ATT&CK® Framework [A],. This framework provides a systematic approach to capture adversarial behaviour targeting cloud environments. Examples of such adversarial behaviours are listed in the Table 1 below.

Table 1. Cloud attacker behaviours

Table 2. Multi-cloud security principles

If telecommunication operators decide to run some of their public telecom network functions in a multi-cloud environment, and specifically in public clouds, the industry will need a set of new standards and new security tools able to regulate the interactions between participating parties. For instance, it seems to be sensible to eliminate access to some restricted areas altogether rather than merely prohibit it. The ETSI standard TS 103 457 “Interface to offload sensitive functions to a trusted domain” [B] provides extra security requirements for public clouds to offer telecommunication operators the option of running public telecom network functions in public clouds.

TS 103 457 introduces a concept of an LTD (Less Trusted Domain) and an MTD (More Trusted Domain) and specifies the TCDI (Trusted Cross-Domain Interface) to standardise secure interactions between them.  The standard defined the following elementary functions of TCDI:

• Connection and session management
• Data and value management
• Transferring cryptography functionality
  • Entropy request
  • Encryption keys request
  • Trusted timestamping
  • Secure archive
  • Secure storage
• Search capabilities

As described in Sec. 1 (Scope) of TS 103 457 it “… specifies a high-level service-oriented interface, as an application layer with a set of mandatory functions, to access secured services provided by, and executed in a More Trusted Domain. The transport layer is out of scope and left to the architecture implementation “. The standard provides extra security for sensitive functions down to individual Virtual Machines or Containers. As such, it is recommended that the relevant components of reference models, reference architecture, reference implementations and reference compliance take notice of this standard and ensure their compatibility, wherever possible.

On the Container side, the NIST Internal Report (NISTIR) 8320A “Hardware-Enabled Security: Container Platform Security Prototype” [D] explains an approach based on hardware-enabled security techniques and technologies for safeguarding container deployments in multi-tenant cloud environments. It also describes a proof-of-concept implementation of the approach – a prototype – that is intended to be a blueprint or template for the general security community.

REFERENCES