Network of the Future Technology DeepdiveExplore the Tech


Comments
Wired

The Profile Management Application: Optimizing DOCSIS® 3.1 Networks

Karthik Sundaresan
Principal Architect, Network Technologies

Jan 14, 2019

Cable operators are broadly deploying DOCSIS 3.1 technology around the world. Operators in North America and Europe have announced gigabit service offerings that use the technology across their footprint. In this blog, I focus on techniques by which operators can boost the capacity and robustness of their existing DOCSIS 3.1 network.

Introducing Toolsets for Optimizing the DOCSIS 3.1 Network

DOCSIS 3.1 specifications introduced orthogonal frequency division multiplexing (OFDM) technology to the access network. In DOCSIS 3.0 and prior technologies, a single-carrier downstream channel (6 MHz) would be configured with just the one modulation order (say, 256-QAM). The DOCSIS 3.1 technology divides a wider (e.g. 192 MHz) channel into a number of small subcarriers spaced at 25 kHz or 50 kHz apart, creating up to 7600 subcarriers. These subcarriers can be individually configured to operate at a different modulation order, anywhere from 256-QAM, to 1024-QAM, all the way to 4096-QAM.

A specific configuration of modulation orders for every subcarrier in the OFDM channel is known as a “profile.” A cable modem termination system (CMTS) can use multiple profiles on a channel; profiles differ in the modulation orders assigned to each subcarrier.

PMA ToolsettingDifferent modems experience the network in different ways. Some modems may have a noisier environment, whereas others see a relatively clean plant. The figure shows an example of the interference seen by a modem in part of the OFDM channel.

Profiles allow the channel to operate at a higher modulation order in clean parts of the spectrum and then go down to more robust modulation orders when there is any interference present. The idea with multiple profiles is to have the CMTS use different profiles for such different groups of cable modems. This allows the operator to reach two goals:

  1. minimize transmission errors on the network and
  2. maximize network capacity.

The same concept applies to upstream OFDMA channels, where one can create upstream profiles for the channel, to increase reliability and throughput.

The Profile Management Application (PMA) Solution

If there are say 4 OFDM channels per port, 10 ports per line-card, and 10 line-cards on a CMTS, then there could be easily over a couple of million subcarriers requiring a specific configuration! This represents an immense configuration challenge which cannot be solved manually. This problem is made more difficult because profiles are not typically “one size fits all,” but rather they need to be tailored to the set of cable modems on the channel, the interference on the channel, and they also need to adjust to plant changes over time. The question then becomes: What is the most efficient method for automating the creation of multiple profiles to maximize robustness and capacity?

I am pleased to announce that CableLabs has met this tremendous challenge with the introduction of the Profile Management Application (PMA). The PMA creates a set of optimal profiles for each channel and assigns profiles to modems. And the PMA accomplishes this dynamically by proactively reading data collected from the network. The PMA can create optimized modulation profiles periodically, as well as backup profiles in case of errors. It can also intelligently decide when to roll out profile changes to the network. A single PMA instance can create and configure profiles for a number of CMTSs in a short period of time and help the operator understand the data capacity of each channel.

How It Works—PMA Architecture and Algorithms

CableLabs released a PMA Technical Report that describes the PMA architecture, use cases, and interface definitions. We also developed data models (YANG) and protocols to facilitate the design of a software application to configure and manage modulation profiles.

Determining the optimal set of modulation profiles to use on a DOCSIS 3.1 channel is a task for intelligent software, given the number of modems, the number of subcarriers and the differences in signal quality across the channel each modem experiences. This CableLabs paper on profile management algorithms describes methods to optimally compute the profiles. The paper also introduces a metric indicating the capacity gain a set of profiles obtains compared with running all traffic at 256-QAM. The key idea is to use clustering algorithms to group modems that have similar signal-to-noise signatures across the channel and then design a profile for each group. The PMA algorithm searches for the best set of profile definitions to maximize the overall capacity and at the same time keeps the individual profiles robust to observable interference patterns.

Some of the information that a PMA needs to compute the profiles includes the channel information from a CMTS, list of modems and network signal quality metrics (mainly, each modem’s per subcarrier RxMER data). Also using performance data (e.g., codeword error rates), a PMA can further refine the profiles.

How It Works—Deployment

How Profile Management Application (PMA) Works

A PMA field deployment would include the following:

  1. Data Collection: A centralized server is needed for modems and CMTSs to upload signal quality data. Operators can use the CableLabs DOCSIS Common Collection Framework (DCCF) to implement this.
  2. PMA: Using the data collected, this application creates optimized profiles per channel. It configures these profiles on the CMTS, assign CMs to profiles and can initiate performance tests.
  3. CMTS: Configures profiles for the individual channels, assigns the modulation profiles to CMs and ultimately sends/receives data using the profiles.
  4. CMs: Cable Modems use the profiles defined and assigned by the CMTS to receive/send data.

What’s Next? PMA Software Available

CableLabs has developed network-deployable software that operators can integrate and use within their DOCSIS 3.1 access network. This PMA solution can use the data generated by the network devices and create optimized OFDM/OFDMA profiles while considering different capabilities supported across CMTS platforms. The software is capable of calculating profiles for a channel with around 200 CMs in ~30 msecs using a single CPU process, so it is easily scalable across an operator’s network. Many operators are testing and planning PMA deployment using this solution framework. PMA is available via CableLabs C3. Contact us to try it out. (Watch for an upcoming video blog on this topic.)

Gains from PMA

A PMA will help increase the throughput per cable modem and maximize network capacity by optimizing the bit-loading of the subcarriers within a channel. Designing profiles around noisy areas in the plant make the system operation more robust. The bandwidth gains in running a well-designed set of profiles can be anywhere from 15% to 40% capacity increase on a channel, compared with running the whole channel at 256-QAM. This can translate to a solid 200 to 400 Mbps extra capacity on each OFDM channel! This enables an operator to match growing bandwidth demands and defer potential node-splits and new equipment costs.

With this profile management technology, operators can realize the full potential of their DOCSIS 3.1 network, by minimizing transmission errors and maximizing the data capacity of the OFDM/OFDMA channels.

Subscribe to our blog to learn more about PMA in the future.


SUBSCRIBE TO OUR BLOG

Comments
DOCSIS

Full Duplex DOCSIS® Technology Gets MAC Layer Support

Karthik Sundaresan
Principal Architect, Network Technologies

Jan 30, 2018

Full Duplex DOCSIS® 3.1 (FDX) is an update and extension to DOCSIS 3.1 specifications that builds on the core Orthogonal Frequency Division Multiplexing (OFDM) technology. This additional set of features significantly increases upstream capacity and allows for the same spectrum to be used as downstream or upstream.

The set of MAC Layer technology changes to support FDX has now been incorporated into the next version of the DOCSIS 3.1 MULPI specification. This supports the PHY layer FDX functionality introduced last October. The new FDX capability and functions are introduced as changes across the MULPI (MAC and Upper Layer Protocols Interface) specification and is now an official part of the specification. This important milestone is the result of a great deal of work by CableLabs members, vendors and employees during the past year, and we take this opportunity to acknowledge their valuable contribution to the cable industry. 

New MAC Layer Functionality Introduced

The MAC layer functionality introduced supports the new FDX operation on the hybrid-fiber-coax (HFC) link.  It is focused on MAC management messaging and operation needed to enable FDX between the CMTS (Cable Modem Termination System) and CM (Cable Modem). This includes FDX channel acquisition/initialization process by a CM, and new processes such as Sounding, Echo Cancellation training, and Resource block assignment.

How it Works – Cable Modem to CMTS Communication

An FDX CMTS will simultaneously receive and transmit in the same FDX spectrum, while FDX CMs can either receive or transmit in the same FDX spectrum. Thus, communication is full duplex from the perspective of the CMTS but frequency division duplex from the perspective of the CM.

The FDX band is divided into sub-bands and the CMTS assigns which sub-band(s) each CM uses for upstream or downstream operation. This is referred to as a resource block assignment (RBA). Different CMs will have different bandwidth demand for both the upstream and downstream directions which can change over time, and FDX allows for the RBA to be changed dynamically.

A sounding method is used to identify groups of CMs, called Interference Groups (IGs), that would interfere with each other if they were allowed to transmit and receive at the same time in a sub-band. IGs are grouped together into a small number of Transmission Groups (TGs), CMs in the same TG either transmit or receive on any given sub-band and time, as signaled by the CMTS in the RBA. CMs from different TGs have enough isolation to transmit and receive at the same time in the same sub-band.

FDX uses a combination of interference cancellation and intelligent scheduling at the CMTS. On the CM, in order to prevent upstream transmissions from interfering with adjacent downstream channels in the FDX band, echo cancellation techniques are used.

Full Duplex DOCSIS® MAC Layer Support

What’s Next for the Full Duplex DOCSIS 3.1 Technology?

As the FDX specifications mature, we will start to see products with full-duplex capabilities over the next year as silicon designs become available and are incorporated into product designs. In an effort to increase the pace of product development, CableLabs has announced a series of FDX Interoperability events, starting in February. These will start from basic node level echo cancellation and gradually progress into full-blown product interoperability.

The collaborative model at CableLabs for developing common cable industry specifications minimizes interoperability issues and gets the best in class features into the specifications. This accelerates time to market for a product, with the operator getting to deployment faster, and ultimately the consumer gets to reap the benefits of the latest technology.

In the meantime, the big wheels of CableLabs specification work continue to turn:

  • We are developing the OSS (Operations Support Systems) changes needed to support FDX.
  • We have initiated work for the changes needed to support FDX within the Remote PHY realm. (FDX assumes a distributed architecture and a plant which supports a node plus zero actives due to the CMTS node echo cancellation functionality)
  • As suppliers build the FDX products, the feedback loop into the FDX specifications remains open and the PHY & MAC specifications continue to be refined.

Full Duplex DOCSIS 3.1 significantly increases upstream capacity on the HFC network, allows flexible splits of upstream and downstream capacity, and enables cable operators to deploy multi-gigabit symmetrical services. With all the innovations being developed by the industry, it is a great time to be a consumer of high-speed data services over cable.

 

FDX MAC Layer Support

Karthik Sundaresan is a Principal Architect at CableLabs responsible for the development and architecture of cable access network technologies. He is primarily involved in the DOCSIS family of technologies and their continued evolution.

Subscribe to our blog to stay current on Full Duplex DOCSIS technology. 

Comments
DOCSIS

New Open Source Initiative at CableLabs

Karthik Sundaresan
Principal Architect, Network Technologies

Feb 10, 2016

Open source software continues to make solid inroads in the world of network technology. There are various open source industry efforts which are becoming de-facto standards that are being adopted by operators and equipment manufacturers (Linux, Apache, OpenStack, Docker etc). Open source leads to free and quick proliferation of good ideas. Collaboration tackles tough problems which may not be solved individually. The open source approach facilitates a rapid prototyping for new technologies and allows improvement on the most important features. It also allows communities to form around a common cause.

CableLabs is increasing its focus on game-changing innovations and accelerating the delivery of unique competitive advantages to the global cable industry with its CableLabs 2.0 initiative. As a part of this focus, CableLabs would like to announce a new major open/community source project for the cable industry. CableLabs and Cisco are initiating a new collaborative project called ‘OpenRPD’ to develop software which can be used by the industry to build a Remote PHY Device. Cisco will be contributing their Remote PHY interface software to this effort, which forms a baseline on which this project will build. (See Cisco's press release.)

Over the past year, CableLabs, along with our member and vendor community, has worked on the different Distributed CCAP Architectures including Remote PHY and Remote MAC-PHY Architectures. The Remote PHY technology allows for an integrated CCAP to be separated into two components: the CCAP Core and the Remote PHY Device (RPD). The RPD allows all the PHY layer components of a CCAP to be moved out as a separate device into the fiber node in the field.

Remote PHY is a big transition in the traditional access network architecture. To meet the deployment needs of the operators, it is imperative we move the ecosystem quickly. The OpenRPD software effort allows us to do this by enabling faster development of RPD products. We believe that open source software is becoming the new way to create specifications, the mantra now is to ‘write code’ not documents.
Word Cloud "Open Source"
A collaborative effort allows for greater reliability in software products, while offering a greater level of security, both of which are important to an RPD platform which will be out in the field. Developing a common code base for some of the basic RPD functions creates a software platform which will minimize interoperability issues between the CCAP-Core and the RPD. It enables companies to focus on their added value and accelerates time to market for a product. This creates a scenario in which everybody wins and the operator gets to deployment of technology faster.

We welcome all interested developers within the CableLabs community to participate in this project. If you would like to participate in the CableLabs OpenRPD Software initiative, please contact me.

May the source be with you.

Karthik Sundaresan is a Principal Architect at CableLabs responsible for the development and architecture of cable access network technologies. He is primarily involved in the DOCSIS family of technologies and their continued evolution.

Comments
DOCSIS

CableLabs® New Remote PHY Specifications expand DOCSIS® Network deployment options

Karthik Sundaresan
Principal Architect, Network Technologies

Jul 7, 2015

DOCSIS technology continues to extend the usefulness of the hybrid fiber coaxial network and increase its global adoption. Distributed Architectures for DOCSIS networks are emerging that provide significant scale advantages and flexible deployment options supporting for both DOCSIS 3.0 and DOCSIS 3.1 networks. Distributed DOCSIS deployments are beginning today in some markets based on the earlier C-DOCSIS specifications.

New Specification Release

CableLabs is documenting several different Distributed CCAP Architectures (including Remote PHY and Remote MAC-PHY) and will release the set of technical reports and specifications throughout this summer.

Last month, CableLabs publicly issued the Remote PHY family of specifications. Theses specifications are also known as MHAv2 as these are an evolution from the original Modular Headend Architecture specifications.

The Remote PHY technology allows for an integrated CCAP to be separated into two components: the CCAP Core and the Remote PHY Device (RPD) and describes the interfaces between them. One of the common locations for an RPD is the optical node device that is located at the junction of the fiber and coax plants, while the CCAP Core stays at the headend. A CCAP core can control and setup data paths with multiple RPDs situated in multiple fiber nodes.

remote-phy-karthik

 

What is Remote PHY?

The Remote PHY technology uses pseudowires between a CCAP Core and a set of RPDs. The CCAP Core contains both a CMTS Core for supporting DOCSIS data transport and an Edge QAM Core for supporting video transport. The CMTS Core contains the DOCSIS MAC (signaling functions, downstream and upstream bandwidth scheduling, and DOCSIS framing) and the upper layer protocols. Remote PHY supports both DOCSIS 3.0 & 3.1 Specifications. The EQAM Core contains all the video processing functions that an EQAM provides today.

The RPD contains mainly PHY related circuitry, such as downstream QAM and OFDM modulators, upstream QAM and OFDM demodulators, together with pseudowire logic needed to connect to the CCAP Core. The RPD platform is a physical layer converter device whose functions are to convert downstream DOCSIS data, MPEG video and out-of-band (OOB) signals received from a CCAP Core over a digital fiber network such as Ethernet or passive optical network (PON) to analog RF for transmission over the coaxial cable; and to convert upstream RF DOCSIS, and OOB signals received over the coaxial cable to digital for transmission over Ethernet or PON to a CCAP Core.

r-phy-docsis-signaling

The CableLabs Remote PHY technology is detailed by six specifications and one technical report (describing the overall architecture) including:

  • The System Specification that describes System level requirements such as initialization sequences and security.
  • The R-DEPI and R-UEPI specifications that describe the downstream and upstream pseudowires and the L2TPv3 control plane.
  • The GCP specification that defines a protocol used for configuration of Remote PHY Devices (RPD).
  • The R-DTI specification that defines the timing interface between the CCAP-Core and RPD.
  • The R-OOB specification that defines support for the SCTE55-1 and 55-2 out of band data for video applications.

What’s Next?

These specifications define the technology to provide guidance to vendors building solutions for the Remote PHY architecture. Vendors have begun architecting ASIC designs, device platforms and software to implement the RPD and CCAP-Core devices. The OSS requirements for managing these devices are also being specified at CableLabs and will be released as an additional specification later this summer. These distributed architectures of course support standard DOCSIS 3.0 and 3.1 modems and gateways no differently than integrated architectures.

The main options under the umbrella of Distributed CCAP Architectures are the Remote PHY and the Remote MAC-PHY technologies. CableLabs’ work is in progress to document the Remote MAC-PHY architecture. This work will culminate in a technical report which will also be released this summer.

Investigating Distributed CCAP Architectures

The work around Distributed CCAP architectures (DCA) is of interest to many CableLabs members in North America, Europe and Asia. Cable operators are investigating DCA for the various gains they bring including:

  • Maximizing DOCSIS 3.1 Channel capacity
  • Simpler operations with digital fiber/Ethernet transport
  • Higher Efficiency of Digital Optics vs. Analog Optics (wavelengths, reach, cost)
  • Helps hub/headend facilities issues around space, power, and cooling as operators move towards Fiber Deep architectures or consider further Node Splits
  • Consistency with FTTx deployments which will include remote architectures for reach and wavelength management
  • Fits with the SDN/NFV initiatives operators are considering across access networks

These integrated & distributed HFC technologies have parallels, and similar features and benefits, to wireless infrastructure architectures such as Macro-cells, small-cells, Distributed Antenna Solutions, and Cloud-RAN with Remote Radio Units. These Distributed CCAP Architectures fit well in different deployment scenarios and all work cohesively together to support the varying capacity and demand in the areas where their deployments provide the best solution.

As operators look to optimize their network deployments in each of their cable plants in each of their markets it will be very interesting to see how and where these distributed CCAP technologies will be deployed in each operator’s HFC networks. It is indeed an exciting time to be working on the access network technologies and being part of the evolution it is going through.

Karthik Sundaresan is a Principal Architect at CableLabs, responsible for the development and architecture of cable access network technologies. He is primarily involved in the DOCSIS family of technologies and their continued evolution.

Comments
DOCSIS

CableLabs Technology Adopted by China’s Cable Operators (C-DOCSIS)

Karthik Sundaresan
Principal Architect, Network Technologies

Sep 4, 2014

DOCSIS® continues to prove its strength, longevity, and relevance as a worldwide access network technology. Last week, CableLabs issued the C-DOCSIS System Specification, completing the suite of specifications known as C-DOCSIS. Cable companies in China are choosing DOCSIS as one of the technologies for offering High-Speed Data and other IP services.

CableLabs has new member companies in China, WASU Digital TV Media Group Co., Shenzhen Topway Video Communication Co., Beijing Gehua CATV Co., and Jiangsu Broadcasting Cable Information Network Corp., who provided leadership in uniting around a common set of requirements specific to the cable networks in China. The Chinese cable operators, along with the State Administration of Radio, Film and Television (SARFT), have been working with CableLabs to develop the technology requirements and architecture for DOCSIS equipment within China.

Cable Modem

 

What is C-DOCSIS?
In 2013, CableLabs published annexes to the DOCSIS 3.0 family of specifications to account for modifications for China. Previously, we had published Euro annexes to document changes necessary for the European cable plant. The new C-DOCSIS System specification defines requirements to support three different distributed CMTS architectures. The DOCSIS 3.0 Annexes and the technology options defined in the System specification are together referenced as 'C-DOCSIS'.

The C-DOCSIS System Spec introduces different options for distributed CMTS architectures, where the CMTS functionality is split between two devices. One is a Coax Media Converter (CMC) device, which bridges the optical and coax domains and resides in the Fiber Node. The other is a CMC controller, which remains at the hub site/head-end. In these architectures the hub site and the fiber node are connected by a digital optical packet link (GE, 10 GE) instead of an analog optical link. The options essentially differ in where the majority of the DOCSIS MAC Layer and PHY Layer functions reside: CMC1 (Remote MAC+PHY) where both the MAC and the PHY are pushed down to the node, CMC2 (Split MAC) where part of the MAC remains at the hub-site and part of the MAC and the PHY are pushed down to the node, and the third option CMC3(Remote PHY) where the MAC stays at the hub-site and only the PHY is moved to the fiber node.

What's Next?
At high level, these developments act as a catalyst in spurring innovations in the CMTS market place, as cable operators across the world are investigating different distributed architectures. These architectures provide myriad benefits for cable operators, including gains in PHY layer performance, hub site space and cooling improvements, and CapEx benefits from using digital optics. CMTS vendors have been actively developing innovative products in the distributed CMTS architecture space that map into one of these architectures.

Consolidating the C-DOCSIS extensions within the main line of CableLabs' DOCSIS specifications ensures that we have one DOCSIS technology that is applicable across the globe, including Europe, North America, and China. CableLabs will continue to bring the C-DOCSIS System spec and all the latest DOCSIS standards towards international standardization through contributions to the European Telecommunications Standards Institute(ETSI). We will also plan interoperability events and other testing to facilitate the growth of this ecosystem. These C-DOCSIS specifications are the first step in engaging our new members in China. It promises to benefit all of our members worldwide by bringing a larger scale to DOCSIS deployments.

Karthik Sundaresan is a Principal Architect at CableLabs, responsible for the development and architecture of cable access network technologies. He is primarily involved in the DOCSIS family of technologies and their continued evolution.

Comments