5. Feature set and Requirements from Infrastructure

A profile Profiles, Profile Extensions & Flavours specifies the configuration of a Cloud Infrastructure node (host or server); Profile Extensions (specialisations) may specify additional configuration. Workloads utilise profiles to describe the configuration of nodes on which they can be hosted to execute on. Workload Flavours provide a mechanism to specify the VM or Pod sizing information to host the workload. Depending on the requirements of the workloads, a VM or a Pod will be deployed as per the specified Flavour information on a node configured as per the specified Profile. Not only do the nodes (the hardware) have to be configured but some of the capabilities also need to be configured in the software layers (such as Operating System and Virtualisation Software). Thus, a Profile can be defined in terms of configuration needed in the software layers, the Cloud Infrastructure Software Profile, and the hardware, the Cloud Infrastructure Hardware Profile.

5.1. Cloud Infrastructure Software profile description

Cloud Infrastructure Software layer is composed of 2 layers, Figure 5.1:

• The virtualisation Infrastructure layer, which is based on hypervisor virtualisation technology or container-based virtualisation technology. Container virtualisation can be nested in hypervisor-based virtualisation

• The host OS layer

Figure 5.1 Cloud Infrastructure software layers

Ref	Cloud Infrastructure Software	Type	Definition/Notes	Capabilities Reference (1)
infra.sw.001	Host Operating System	<value>	Values such as Ubuntu20.04, Windows 10 Release #, etc.	e.cap.021
infra.sw.002	Virtualisation Infrastructure Layer	<value>	Values such as KVM, Hyper-V, Kubernetes, etc.	e.cap.022

(1) Reference to the capabilities defined in Infrastructure Capabilities, Measurements and Catalogue.

For a host (compute node or physical server), the virtualisation layer is an abstraction layer between hardware components (compute, storage, and network resources) and virtual resources allocated to a VM or a Pod. Figure 5.2 represents the virtual resources (virtual compute, virtual network, and virtual storage) allocated to a VM or a Pod and managed by the Cloud Infrastructure Manager.

Figure 5.2 Cloud Infrastructure Virtual resources

A Cloud Infrastructure Software Profile is a set of features, capabilities, and metrics offered by a Cloud Infrastructure software layer and configured in the software layers (the Operating System (OS) and the virtualisation software (such as hypervisor)). Figure 5.3 depicts a high level view of the Basic and High Performance Cloud Infrastructure Profiles.

Figure 5.3 Cloud Infrastructure Software Profiles

The following sections detail the Cloud Infrastructure Software Profile capabilities per type of virtual resource.

5.1.1. Virtual Compute Profiles

Table 5-1 and Table 5-2 depict the features related to virtual compute.

Reference	Feature	Type	Description	Capabilities Reference
infra.com.cfg.001	CPU allocation ratio	<value>	Number of virtual cores per physical core.	i.cap.016
infra.com.cfg.002	NUMA alignment	Yes/No	Support of NUMA at the Host OS and virtualisation layers, in addition to hardware.	e.cap.007
infra.com.cfg.003	CPU pinning	Yes/No	Binds a vCPU to a physical core or SMT thread. Configured in OS and virtualisation layers.	e.cap.006
infra.com.cfg.004	Huge pages	Yes/No	Ability to manage huge pages of memory. Configured in OS and virtualisation layers.	i.cap.018
infra.com.cfg.005	Simultaneous Multithreading (SMT)	Yes/No/Optional	Allows multiple execution threads to be executed on a single physical CPU core. Configured in OS, in addition to the hardware.	e.cap.018

Table 5-1: Virtual Compute features.

Reference	Feature	Type	Description	Capabilities Reference
infra.com.acc.cfg .001	IPSec Acceleration	Yes/No/Optional	IPSec Acceleration	e.cap.008
infra.com.acc.cfg .002	Transcoding Acceleration	Yes/No/Optional	Transcoding Acceleration	e.cap.010
infra.com.acc.cfg .003	Programmable Acceleration	Yes/No/Optional	Programmable Acceleration	e.cap.011
infra.com.acc.cfg .004	GPU	Yes/No/Optional	Hardware coprocessor	e.cap.014
infra.com.acc.cfg .005	FPGA/other Acceleration H/W	Yes/No/Optional	Non-specific hardware. These Capabilities generally require hardware-dependent drivers be injected into workloads.	e.cap.016

Table 5-2: Virtual Compute Acceleration features.

5.1.2. Virtual Storage Profiles

Table 5-3 and Table 5-4 depict the features related to virtual storage.

Reference	Feature	Type	Description
infra.stg.cfg.001	Catalogue Storage Types	Yes/No	Support of Storage types described in the catalogue
infra.stg.cfg.002	Storage Block	Yes/No
infra.stg.cfg.003	Storage with replication	Yes/No
infra.stg.cfg.004	Storage with encryption	Yes/No

Table 5-3: Virtual Storage features.

Reference	Feature	Type	Description
infra.stg.acc.cfg.001	Storage IOPS oriented	Yes/No
infra.stg.acc.cfg.002	Storage capacity oriented	Yes/No

Table 5-4: Virtual Storage Acceleration features.

5.1.3. Virtual Networking Profiles

Table 5-5 and Table 5-6 depict the features related to virtual networking.

Reference	Feature	Type	Description	Capabilities Reference
infra.net.cfg.001	Connection Point interface IO virtualisation	e.g. virtio1.1
infra.net.cfg.002	Overlay protocol	Protocols	The overlay network encapsulation protocol needs to enable ECMP in the underlay to take advantage of the scale-out features of the network fabric.
infra.net.cfg.003	NAT	Yes/No	Support of Network Address Translation
infra.net.cfg.004	Security Groups	Yes/No	Set of rules managing incoming and outgoing network traffic
infra.net.cfg.005	Service Function Chaining	Yes/No	Support of Service Function Chaining (SFC)
infra.net.cfg.006	Traffic patterns symmetry	Yes/No	Traffic patterns should be optimal, in terms of packet flow. North-south traffic shall not be concentrated in specific elements in the architecture, making those critical choke-points, unless strictly necessary (i.e. when NAT 1:many is required).

Table 5-5: Virtual Networking features.

Reference	Feature	Type	Description	Capabilities Reference
infra.net.acc.cfg.001	vSwitch optimisation	Yes/No and SW Optimisation	e.g. DPDK.	e.cap.019
infra.net.acc.cfg.002	SmartNIC (for HW Offload)	Yes/No	HW Offload	e.cap.015
infra.net.acc.cfg.003	Crypto acceleration	Yes/No		e.cap.009
infra.net.acc.cfg.004	Crypto Acceleration Interface	Yes/No

Table 5-6: Virtual Networking Acceleration features.

5.1.4. Security

See Chapter 7 Security.

5.1.5. Platform Services

This section details the services that may be made available to workloads by the Cloud Infrastructure.

Reference	Feature	Type	Description
infra.svc.stg.001	Object Storage	Yes/No	Object Storage Service (e.g S3-compatible)

Table 5-7: Cloud Infrastructure Platform services.

Platform Service Category	Platform Service Examples
Data Stores/Databases	Ceph, etcd, MongoDB, Redis
Streaming and Messaging	Apache Kafka, Rabbit MQ
Load Balancer and Service Proxy	Envoy, Istio, NGINX
Service Mesh	Envoy, Istio
Security & Compliance	Calico, cert-manager
Monitoring	Prometheus, Grafana (for Visualisation), Kiali (for Service Mesh)
Logging	Fluentd, ElasticSearch (Elastic.io, Open Distro), ELK Stack (Elasticsearch, Logstash, and Kibana)
Application Definition and Image Build	Helm
CI/CD	Argo, GitLab, Jenkins
Ingress/Egress Controllers	Envoy, Istio, NGINX
Network Service	CoreDNS, Istio
Coordination and Service Discovery	CoreDNS, etcd, Zookeeper
Automation and Configuration	Ansible
Key Management	Vault
Tracing	Jaeger

Table 5-7a: Service examples.

5.1.5.1. Platform Services - Load Balancer Requirements

The table below specifies a set of requirements for the Load Balancer platform service.

Reference	Requirement	Notes
pas.lb.001	The Load Balancer must support workload resource scaling
pas.lb.002	The Load Balancer must support resource resiliency
pas.lb.003	The Load Balancer must support scaling and resiliency in the local environment	Local environment: within a subnet, tenant network, Availability Zone of a cloud, …
pas.lb.004	The Load Balancer must support OSI Level 3/4 load-balancing	OSI Level 3 load-balancing decision on the source and destination IP addresses and OSI Level 4 TCP port numbers.
pas.lb.005	The Load Balancer must, at a minimum, support round-robin load-balancing
pas.lb.006	The Load Balancer must create event logs with the appropriate severity levels (catastrophic, critical, …)
pas.lb.007	The Load Balancer must support monitoring of endpoints
pas.lb.008	The Load Balancer must support Direct Server Return (DSR)	Other modes OK as well, but DSR should always be supported
pas.lb.009	The Load Balancer must stateful TCP load-balancing
pas.lb.010	The Load Balancer must support UDP load-balancing
pas.lb.011	The Load Balancer must support load-balancing and correct handling of fragmented packets
pas.lb.012	The Load Balancer may support state-full SCTP load-balancing
pas.lb.013	The Load Balancer may support state-full M-TCP load-balancing
pas.lb.014	The Load Balancer may support Level 7 load balancing	OSI Level 7 (application characteristics based) should support HTTP and HTTPS
pas.lb.0156 | The L7 Load Balancer may support HTTP2
pas.lb.016	The L7 Load Balancer may support HTTP3
pas.lb.017	The L7 Load Balancer may support QUIC

Table 5-7b: Platform Services - Load Balancer Requirements.

5.1.5.2. Platform Services - Log Management Service (LMS)

The table below specifies a set of requirements for the Log Management Service (LMS).

Reference	Requirement	Notes
pas.lms.001	LMS must support log management from multiple, distributed sources
pas.lms.002	LMS must manage log rotation at configurable time periods
pas.lms.003	LMS must manage log rotation at configurable log file status (%full)
pas.lms.004	LMS must manage archival and retention of logs for configurable time periods by different log types
pas.lms.005	LMS must ensure log file integrity (no changes, particularly changes that may affect the completeness, consistency, and accuracy including event times, of the log file content)	Covered by req.sec.mon.005: “The Prod-Platform and NonProd-Platform must secure and protect all logs (containing sensitive information) both in-transit and at rest.”
pas.lms.006	LMS must monitor log rotation and log archival processes
pas.lms.007	LMS must monitoring the logging status of all log sources
pas.lms.008	LMS must ensure that each logging host’s clock is synched to a common time source
pas.lms.009	LMS must support reconfiguring of logging as needed based on policy changes, technology changes, and other factors
pas.lms.010	LMS must support the documenting and reporting of anomalies in log settings, configurations, and processes
pas.lms.011	LMS must support the correlating of entries from multiple logs that relate to the same event
pas.lms.012	LMS must support the correlating of multiple log entries from a single source or multiple sources based on logged values (e.g., event types, timestamps, IP addresses)
pas.lms.013	LMS should support rule-based correlation

Table 5-7c: Platform Services - Log Management Service (LMS) Requirements.

5.1.5.3. Platform Services - Monitoring Service Requirements

The table below specifies a set of requirements for the Monitoring service (aka monitoring system).

Reference	Requirement	Notes
pas.mon.001	The Monitoring service must be able to collect data generated by or collected from any resource (physical and virtual infrastructure, application, network, etc.)	Capabilities to monitor applications, services, operating systems, network protocols, system metrics and infrastructure components
pas.mon.002	The Monitoring service must be able to aggregate collected data
pas.mon.003	The Monitoring service must be able to correlate data from different systems
pas.mon.004	The Monitoring service must be able to perform at least one of active or passive monitoring
pas.mon.005	The Monitoring service must support configuration of thresholds, outside of which the resource cannot function normally, for alert generation
pas.mon.006	The Monitoring service must support configuration of alert notification medium (email, SMS, phone, etc.)
pas.mon.007	The Monitoring service must support configurable re-alerting after a configurable period of time if the metric remains outside of the threshold
pas.mon.008	The Monitoring service must support configurable alert escalations
pas.mon.009	The Monitoring service must support alert acknowledgments by disabling future alerting of the same resource/reason
pas.mon.010	The Monitoring service must support selective enabling and disabling of alerts by resource, category of resources, time periods.
pas.mon.011	The monitoring service must publish its APIs for programmatic invocation of all monitoring service functions
pas.mon.012	The monitoring service must itself be monitored through a logging service
pas.mon.013	The Monitoring service should be implemented for high availability to ensure non-stop monitoring of critical infrastructure components
pas.mon.014	The Monitoring service should run as separately from production services
pas.mon.015	Failure of the system being monitored should not cause a failure in the monitoring service
pas.mon.016	An inoperative monitoring service should not generate alerts about the monitored system
pas.mon.017	The monitoring service should provide a consolidated view of the entire monitored infrastructure	View: dashboard or report

Table 5-7d: Platform Services - Monitoring Service Requirements.

5.2. Cloud Infrastructure Software Profiles features and requirements

This section will detail Cloud Infrastructure Software Profiles and associated configurations for the 2 types of Cloud Infrastructure Profiles: Basic and High Performance.

5.2.1. Virtual Compute

Table 5-8 depicts the features and configurations related to virtual compute for the two (2) Cloud Infrastructure Profiles.

Reference	Feature	Type	Basic	High Performance
infra.com.cfg.001	CPU allocation ratio	<value>	N:1	1:1
infra.com.cfg.002	NUMA alignment	Yes/No	N	Y
infra.com.cfg.003	CPU pinning	Yes/No	N	Y
infra.com.cfg.004	Huge pages	Yes/No	N	Y
infra.com.cfg.005	Simultaneous Multithreading (SMT)	Yes/No/Optional	Y	Optional

Table 5-8: Virtual Compute features and configuration for the 2 types of Cloud Infrastructure Profiles.

Table 5-9 lists the features related to compute acceleration for the High Performance profile. The table also lists the applicable Profile Extensions and Extra Specs that may need to be specified.

Reference	Feature	Profile-Extensions	Profile Extra Specs
infra.com.acc.cfg.001	IPSec Acceleration	Compute Intensive GPU
infra.com.acc.cfg.002	Transcoding Acceleration	Compute Intensive GPU	Video Transcoding
infra.com.acc.cfg.003	Programmable Acceleration	Firmware-programmable adapter	Accelerator
infra.com.acc.cfg.004	GPU	Compute Intensive GPU
infra.com.acc.cfg.005	FPGA/other Acceleration H/W	Firmware-programmable adapter

Table 5-9: Virtual Compute Acceleration features.

5.2.2. Virtual Storage

Table 5-10 and Table 5-11 depict the features and configurations related to virtual storage for the two (2) Cloud Infrastructure Profiles.

Reference	Feature	Type	Basic	High Performance
infra.stg.cfg.001	Catalogue storage Types	Yes/No	Y	Y
infra.stg.cfg.002	Storage Block	Yes/No	Y	Y
infra.stg.cfg.003	Storage with replication	Yes/No	N	Y
infra.stg.cfg.004	Storage with encryption	Yes/No	Y	Y

Table 5-10: Virtual Storage features and configuration for the two (2) profiles.

Table 5-11 depicts the features related to Virtual storage Acceleration

Reference	Feature	Type	Basic	High Performance
infra.stg.acc.cfg.001	Storage IOPS oriented	Yes/No	N	Y
infra.stg.acc.cfg.002	Storage capacity oriented	Yes/No	N	N

Table 5-11: Virtual Storage Acceleration features.

5.2.3. Virtual Networking

Table 5-12 and Table 5-13 depict the features and configurations related to virtual networking for the 2 types of Cloud Infrastructure Profiles.

Reference	Feature	Type	Basic	High Performance
infra.net.cfg.001	Connection Point interface	IO virtualisation	virtio1.1	virtio1.1*
infra.net.cfg.002	Overlay protocol	Protocols	VXLAN, MPLSoUDP, GENEVE, other	VXLAN, MPLSoUDP, GENEVE, other
infra.net.cfg.003	NAT	Yes/No	Y	Y
infra.net.cfg.004	Security Group	Yes/No	Y	Y
infra.net.cfg.005	Service Function Chaining	Yes/No	N	Y
infra.net.cfg.006	Traffic patterns symmetry	Yes/No	Y	Y

Table 5-12: Virtual Networking features and configuration for the 2 types of SW profiles.

Note: * might have other interfaces (such as SR-IOV VFs to be directly passed to a VM or a Pod) or NIC-specific drivers on guest machines transiently allowed until mature enough solutions are available with a similar efficiency level (for example regarding CPU and energy consumption).

Reference	Feature	Type	Basic	High Performance
infra.net.acc.cfg.001	vSwitch optimisation (DPDK)	Yes/No and SW Optimisation	N	Y
infra.net.acc.cfg.002	SmartNIC (for HW Offload)	Yes/No/Optional	N	Optional
infra.net.acc.cfg.003	Crypto acceleration	Yes/No/Optional	N	Optional
infra.net.acc.cfg.004	Crypto Acceleration Interface	Yes/No/Optional	N	Optional

Table 5-13: Virtual Networking Acceleration features.

5.3. Cloud Infrastructure Hardware Profile description

The support of a variety of different workload types, each with different (sometimes conflicting) compute, storage, and network characteristics, including accelerations and optimizations, drives the need to aggregate these characteristics as a hardware (host) profile and capabilities. A host profile is essentially a “personality” assigned to a compute host (also known as physical server, compute host, host, node, or pServer). The host profiles and related capabilities consist of the intrinsic compute host capabilities (such as number of CPU sockets, number of cores per CPU, RAM, local disks and their capacity, etc.), and capabilities enabled in hardware/BIOS, specialised hardware (such as accelerators), the underlay networking, and storage.

This chapter defines a simplified host, profile and related capabilities model associated with each of the different Cloud Infrastructure Hardware Profile and related capabilities; the two Profiles, Profile Extensions & Flavours (aka host profiles, node profiles, hardware profiles) and some of their associated capabilities are shown in Figure 5.4.

Figure 5.4 Cloud Infrastructure Hardware Profiles and host associated capabilities

The profiles can be considered to be the set of EPA-related (Enhanced Performance Awareness) configurations on Cloud Infrastructure resources.

Note: In this chapter we shall not list all of the EPA-related configuration parameters.

A given host can only be assigned a single host profile; a host profile can be assigned to multiple hosts. In addition to the host profile, Profiles and Workload Flavours and additional capability specifications for the configuration of the host can be specified. Different Cloud Service Providers (CSP) may use different naming standards for their host profiles. For the profiles to be configured, the architecture of the underlying resource needs to be known.

Ref	Cloud Infrastructure Resource	Type	Definition/Notes	Capabilities Reference
infra.hw.001	CPU Architecture	<value>	Values such as x64, ARM, etc.	e.cap.020

The host profile properties are specified in the following sub-sections. The following diagram (Figure 5.5) pictorially represents a high-level abstraction of a physical server (host).

Figure 5.5 Generic model of a compute host for use in Host Profile configurations

5.4. Cloud Infrastructure Hardware Profiles features and requirements.

The configurations specified in here will be used in specifying the actual hardware profile configurations for each of the Cloud Infrastructure Hardware Profiles depicted in Figure 5-4.

5.4.1. Compute Resources

Reference	Feature	Description	Basic	High Performance
infra.hw.cpu.cfg.001	Minimum number of CPU sockets	Specifies the minimum number of populated CPU sockets within each host (*)	2	2
infra.hw.cpu.cfg.002	Minimum number of cores per CPU	Specifies the number of cores needed per CPU (*)	20	20
infra.hw.cpu.cfg.003	NUMA alignment	NUMA alignment enabled and BIOS configured to enable NUMA	N	Y
infra.hw.cpu.cfg.004	Simultaneous Multithreading (SMT) SMT enabled that allows each core to work multiple streams of data simultaneously	Y	Optional

Table 5-14: Minimum sizing and capability configurations for general purpose servers.

(*) Please note that these specifications are for general purpose servers normally located in large data centres. Servers for specialised use with the data centres or other locations, such as at edge sites, are likely to have different specifications.

5.4.1.1. Compute Acceleration Hardware Specifications

Reference	Feature	Description	Basic	High Performance	Capabilities Reference
infra.hw.cac.cfg.001	GPU	GPU	N	Optional	e.cap.014
infra.hw.cac.cfg.002	FPGA/other Acceleration H/W	HW Accelerators	N	Optional	e.cap.016

Table 5-15: Compute acceleration configuration specifications.

5.4.2. Storage Configurations

Reference	Feature	Description	Basic	High Performance
infra.hw.stg.hdd.cfg.001*	Local Storage HDD	Hard Disk Drive
infra.hw.stg.ssd.cfg.002*	Local Storage SSD	Solid State Drive	Recommended	Recommended

Table 5-16: Storage configuration specification.

Note: *This specified local storage configurations including # and capacity of storage drives.

5.4.3. Network Resources

5.4.3.1. NIC configurations

Reference	Feature	Description	Basic	High Performance
infra.hw.nic.cfg.001	NIC Ports	Total number of NIC Ports available in the host	4	4
infra.hw.nic.cfg.002	Port Speed	Port speed specified in Gbps (minimum values)	10	25

Table 5-17: Minimum NIC configuration specification.

5.4.3.2. PCIe Configurations

Reference	Feature	Description	Basic	High Performance
infra.hw.pci.cfg.001	PCIe slots	Number of PCIe slots available in the host	8	8
infra.hw.pci.cfg.002	PCIe speed		Gen 3	Gen 3
infra.hw.pci.cfg.003	PCIe Lanes		8	8

Table 5-18: PCIe configuration specification.

5.4.3.3. Network Acceleration Configurations

Reference	Feature	Description	Basic	High Performance	Capabilities Reference
infra.hw.nac.cfg.001	Crypto Acceleration	IPSec, Crypto	N	Optional	e.cap.009
infra.hw.nac.cfg.002	SmartNIC	offload network functionality	N	Optional	e.cap.015
infra.hw.nac.cfg.003	Compression		Optional	Optional
infra.hw.nac.cfg.004	SR-IOV over PCI-PT	SR-IOV	N	Optional	e.cap.013

Table 5-19: Network acceleration configuration specification.