Sarat Puthenpura, Chief Architect – Open RAN; Aether (beforehand ONF) can be talking at this yr’s Open RAN International Discussion board. Learn extra and register right here.
Vitality prices for telecom operators will be as excessive as 10% of their whole working expenditure. That is very true as they roll out 5G. Despite the fact that the 5G-NR requirements are extra power environment friendly than their predecessors, the general energy consumption of 5G networks will be considerably increased than earlier technology networks resulting from plenty of compounding components.
Huge MIMO and better MIMO modes, that are extremely useful to assist cell broadband at excessive frequency bands (FR2), require extra energy. Moreover, with cell edge computing, the variety of knowledge facilities will enhance and computationally demanding use instances will additional intensify power utilization. One other side of 5G is the upper working bandwidth for every channel, which has elevated from 20 MHz (LTE) to as excessive as 400 MHz. The sub-division of bigger bandwidths into Bandwidth Components (BWPs), which is geared toward decreasing energy consumption of units, will trigger incremental power consumption for the bottom stations. Thus total, the power consumption in cell networks is anticipated to go up considerably as 5G is rolled out. With all this, cell operators are underneath intense stress to deploy clever options to optimize the power consumption of their cell networks.
Vitality effectivity and power financial savings use instances have been recognized as a high precedence by cell community operators and this drives ongoing standardization discussions within the O-RAN Alliance and 3GPP. Vitality effectivity testing, metrics and estimation fashions present the muse for options for monitoring and optimizing power utilization. Current specs and strategies within the context of Open RAN are being improved and up to date. The general logical steps listed below are easy and intuitive: (1) measure power consumption precisely with adequate granularity; (2) create KPIs and fashions to seize the traits of power consumption; and (3) use the KPIs and fashions to optimize power consumption.
The issue with the present panorama of power saving use instances and related optimization methods is that they’re at Step 3 of the above-mentioned logical steps, considerably bypassing Steps 1 and a pair of. For instance, cell on/off approaches (turning cells on/off based mostly on load) do make sense to avoid wasting power generally, since one may briefly flip some capability off underneath low load situations if a sure stage of high quality of service (QoS) will be maintained. Nonetheless, doing the identical and comparable operations based mostly on quantifiable measures of power effectivity and related fashions makes extra sense, as a result of totally different methods and approaches are influenced closely by community configurations, working situations and efficiency necessities.
Advancing power financial savings capabilities
An open-source anchored effort has emerged to pursue development of power financial savings use instances.
The SMaRT-5G initiative — initially began by Open Networking Basis (ONF) — has been constructing a collection of PoCs that exhibit progressively superior power financial savings strategies on each Open-RAN and conventional RAN architectures. The phased collection of PoCs is to progressively ship energy utilization fashions, APIs and open-source software program, with the objective of accelerating MNO adoption and trade influence. We’ve got efficiently demonstrated two SMaRT-5G PoCs within the latest previous. The primary was at Fyuz in October 2023 and the second befell on the NTIA-hosted RIC discussion board in March 2024. The important thing collaborators of those PoCs had been Rimedo Labs, Tietoevry, Intel Labs, Viavi Options and i14y Labs. Extra PoCs are on the best way.
Lately, the Nationwide Telecommunications and Data Administration (NTIA) of the U.S. Division of Commerce has graciously supported the SMaRT-5G initiative with a analysis grant via their Public Wi-fi Provide Chain Innovation Fund (PWSCIF). The $2million grant is funding SMaRT-5G in collaboration with Rutgers College’s WINLAB to analysis, develop and validate correct and efficient check strategies and to create metrics and fashions to measure the power effectivity of 5G community elements in addition to the effectiveness of end-to-end Open RAN based mostly power optimization methods. The main focus is on discovering a minimal set of parameters and eventualities wanted to realize efficient outcomes, within the context of Open RAN networks using a multi-vendor, disaggregated structure with bodily, digital and cloud-native community features. Experimental analysis is being carried out as a collaborative effort between members of Aether and the SMaRT-5G undertaking and Rutgers WINLAB, leveraging WINLAB’s lab infrastructure and lab’s synergy with different multi-vendor Open RAN tasks.
The anticipated end result of this R&D effort is to develop and validate: (a) modern efficient measurements for power consumption of assorted RAN and core elements; (b) power consumption metrics, KPIs and APIs to be supported by RAN and core gear; (c) power consumption fashions which can be utilized in simulation and analytics research; and (d) strategies to evaluate end-to-end power effectivity optimization utilizing totally different algorithms and devoted purposes.
For power consumption measurements and modeling, we have to take into account each bodily features (similar to RU) in addition to cloudified community features (similar to CU, DU, RIC and x/rApps). Understanding the interaction amongst these community features underneath numerous site visitors and environmental situations affecting power consumption is essential.
The Open RAN requirements/testing group is especially related to this undertaking since they’ve the scope for specifying the metrics and check methodology. Our goal is to be actively engaged within the course of. We’ve got designed our plan based mostly on present and trending requirements and goal to contribute our findings again to the identical course of.
Vitality effectivity testing — Platform and methodology
The platform we have now created for this objective known as POET: Platform for O-RAN Vitality Effectivity Testing and its purposeful structure is proven within the determine beneath.
POET Structure (Platform for O-RAN Vitality Effectivity Testing)
POET contains naked steel servers in addition to a Kubernetes cluster for the O-DU, O-CU and Core. It contains bodily O-RUs in addition to SDR-based RUs.
The Kubernetes system contains near-RT RIC, Kepler energy monitoring and assist for SDR RU and simulated RAN. The testbed contains an Amarisoft UE-emulator and modem/cellphone UEs. Energy equipped to all elements is monitored utilizing good PDUs (Server Know-how PRO3X and STV-6521V.) Energy and efficiency metrics are captured and exported to Prometheus and displayed on Grafana dashboards. POET closely leverages open-source elements to implement end-to-end O-RAN eventualities. Open-source software program is free and overtly accessible, thus enabling groups to duplicate leads to a standard reference framework. This aligns with our undertaking goal of contributing our outcomes to affect the trade. The open supply software program we’re utilizing and plan to make use of is from Aether,OAI, srsRAN and O-RAN Software program Group.
Kubernetes servers are Dell R730xd (Intel Xeon E5-2680 v4 2.4GHz, 56 CPUs) whereas the bare-metal servers are Dell R740 (Intel Xeon Gold 6226 2.70GHz, 24 CPUs) and Penguin Relion 1900 (Intel Xeon E5-2620 v4 2.10GHz, 32 CPUs).
Our broad goal is to analysis the very best methodology for measuring power consumption, power effectivity and power financial savings. Whereas power measurement shouldn’t be a brand new area, there may be comparatively little out there by way of standardized methodology, particularly when addressing cloudified NFs. Our goal is to find out the check methodology fairly than examine the effectivity and efficiency of assorted gear. We intend to discover totally different strategies as a part of our work, together with:
(a) Testing underneath totally different person site visitors and cargo and radio eventualities (e.g., frequency bands, channel bandwidth, path loss, MIMO modes, and so forth.)
(b)Testing several types of O-RU (e.g., indoor/outside,) and server architectures (e.g., with and with out accelerators)
(c) Automating the gathering of energy and efficiency metrics and check metadata, and reporting of outcomes
Measuring the facility and power consumption is on the coronary heart of the power effectivity testing and validation of power optimization options. We’ve got adopted totally different approaches for energy measurement:
(a) PDU: The facility, present and voltage equipped to PNFs and servers is obtained by recurrently querying the Energy Distribution Items (PDUs) (nominally each 10 secs), and exporting this knowledge to Prometheus/Grafana. This offers a important ground-truth measurement of energy consumption.
(b) IPMI: Energy and setting variables reported by the servers are additionally monitored utilizing the Clever Platform Administration Interface (IPMI). Queries are made to the server Baseboard Administration Controller (BMC) and exported to Prometheus/Grafana.
(c) Scaphandre: We deployed the Scaphandre open-source power monitoring performance on bare-metal servers working CU/DU software program. Scaphandre measures course of utilization (based mostly on Intel Operating Common Energy Restrict (RAPL)) and estimates the facility consumption utilizing an estimation mannequin.
(d) Kepler: The Kubernetes deployment within the testbed makes use of Kepler exporters on every server sending energy metrics to Prometheus/Grafana. Kepler is an open-source power monitoring performance for Kubernetes methods. It measures node and container utilization (based mostly on Intel RAPL) and estimates the facility consumption utilizing an estimation mannequin.
For power effectivity, we basically have a high-level relationship given by desired community efficiency divided by power consumed within the related portion of the community. We plan to measure power consumption (denominator of the ratio) utilizing the multi-pronged strategy described earlier than. The numerator, desired community efficiency, is a loaded time period and relying on the context, its definition can fluctuate. For instance, MNOs broadly use throughput and cell-edge efficiency in addition to accessibility and retainability to measure efficiency. Our plan is to gather efficiency measures similar to variety of UEs / RRC connections per cell, DL and UL PRB utilization per cell, DL and UL throughput (Mbps) per cell and per UE, DL and UL knowledge quantity (bytes) per cell and per UE, latency per UE and mixture throughout UEs per cell, MCS worth per UE and mixture throughout cell, and traits of UE site visitors combine – and we are going to discover correlations which can be related to power effectivity.
Vitality effectivity testing and power optimization monitoring require correct and handy efficiency KPI measurements. We’ve got the next approaches for KPI metrics:
(a) E2E: We use iPerf utility end-to-end to acquire the info quantity, throughput and latency of a session between the UE and a check server within the packet knowledge community.
(b) O1-based KPI: We use an “O1-like” strategy utilizing a telnet-based answer that gives uplink/downlink all through and downlink PRB load for DU. It is a preliminary model of ongoing open-source work to implement an O1 interface for CU/DU.
(c) E2-based KPI: The E2 interface from the O-CU and O-DU to the Close to-RT RIC and xApps can present a number of KPIs utilizing the E2-SM KPM. Our Kubernetes-based O-RAN system features a KPM xApp on Close to-RT RIC with an E2 interface to the CU/DU.
We plan to make use of our check outcomes to develop fashions to estimate the facility consumption of O-RAN networks, together with:
1) Fashions present estimates based mostly on out there power and efficiency metrics when a given deployed system could not assist all the newest self-reported power metrics
2) Fashions which will be calibrated with preliminary lab-based checks of apparatus that are later deployed within the community
3) Fashions to offer enter for rApp/xApp power saving purposes
It is very important word that POET leverages open-source elements to implement end-to-end O-RAN eventualities. Open-source software program is free and overtly accessible, thus enabling groups to duplicate leads to a standard reference framework. This aligns with our undertaking goal of contributing our outcomes to affect the trade.
It’s also notable that this undertaking is getting collaboration assist from BubbleRAN and Cognizant with extra corporations within the pipeline.
Concluding remarks
POET is operational now, and we have now began getting preliminary measurements from the platform. These measurements are in-line with our expectations and plan. Evaluation, interpretation and modeling actions are progressing together with the enlargement and enhancements of POET. This can be mentioned in a future article.
Sarat Puthenpura, Chief Architect – Open RAN; Aether (beforehand ONF) can be talking at this yr’s Open RAN International Discussion board. Learn extra and register right here.