‘Best effort’ should be considered as critical for the telco of tomorrow

Many in the industry shudder at the concept of ‘best effort’, but this should change in future when latency can be billed as a business case.

The concept of ‘best effort’ is perceived in different ways partly reliant on whether you consider the glass to be half-full or half-empty. Those of a pessimistic nature will hear the term ‘best effort’ and decide those sitting on the other side of the table are unreliable. The more positive side of the population will consider their partner as a person who will try their utmost to provide the best possible service.

This conflict is unavoidable as it is embedded in the personality of the individual, but according to Crystal Web Founder Paul Hjul (pictured) embracing the concept of ‘best effort’ is critical for success in tomorrow’s digital economy.

Crystal Web is much more of a South African challenger ISP than anything else, having only launched commercial services in 2014. It bills itself as a business which operates differently and Hjul certainly lives up to that reputation in his public appearances and thoughts on the industry. With Crystal Web being free from many of legacy restraints which plague traditional telcos, and Hjul’s differentiating mindset, the company is free to operate in a unique manner. This is partly where Hjul’s thoughts on ‘best effort’ are derived from.

To validate the best effort approach, telcos have to appreciate there are two ways to deliver connectivity in the future. There is the commoditized delivery of the internet, and value added services with enhanced connectivity. One has a dumb pipe, while the other has a smart one which can be glorified with the network slicing euphoria. This is where ‘best effort’ should be applied, and is built on the demand for latency.

Many telcos want to avoid the dumb pipe, but why should they? A dumb pipe can effectively deliver on data demands for the vast majority of consumer cases, many of which are not reliant on latency. If it takes an extra five seconds for a WhatsApp message to be delivered, or an extra two seconds for a website to load, it’s not the end of the world. These are scenarios which can be tolerated, and the business case for the dumb pipe and ‘best effort’ can be realised. Building this dumb pipe and effectively utilizing the internet for this segment is perfectly acceptable, and allows telcos to concentrate on other aspects of the business.

The smart pipe is where money can be made. Telcos can work with enterprise organizations who cannot use the public internet for a variety of reasons, or on consumer usecases which require low latency, through the smart pipe to ensure enhanced connectivity. This is where latency plays a factor, but for every nine which is added onto the reliability and resiliency, more money will have to be charged. The usecases here are relatively obvious, autonomous cars for instance or transmission of sensitive data, but each have additional demands on latency, security and reliability.

According to Hjul, there is nothing wrong with a dumb pipe, as long as it is built smartly, but the business case for tomorrow’s telcos has to be established through separating the various usecases. Dumb is only as dumb as the business case which underpins it.

Cutting through the hype of 5G’s promises

Telecoms.com periodically invites expert third parties to share their views on the industry’s most pressing issues. In this piece R. Ezhirpavai, VP of Technology at Aricent, looks at the myths around 5G use cases, including ultra-low latency – something both vendors and operators have been keen to promote. 

The 5G rollout is gathering momentum. Big names in the telco space are confidently demonstrating their nascent next-generation technology, with Qualcomm, for example, recently teasing its first 5G mobile platform, and LG and Sprint revealing plans to deliver the first 5G handset in the first half of 2019. As the number of announcements grow, so too does the level of discussion around what 5G entails.

Ultra-low latency is often heralded as a key benefit by both vendors and operators, particularly in enabling connected cars and automation technology, with scalability and cost-effectiveness regularly cited as additional advantages. Before we get too excited, however, let’s take a step back and view these claims with an objective eye, and examine the truth behind the hype.

Shiny connected cars

Much has been written about how the future of connected cars will depend on 5G connectivity. These vehicles rely on a number of applications and they all require ongoing maintenance. For example, applications to collect telemetry data, gather Software-Over-The-Air (SOTA) and Firmware-Over-The-Air (FOTA) updates. While it is possible to run these applications over existing 4G networks, 5G will be of course be more effective due to the larger capacity it offers at higher speeds.

5G millimetre wave antennae with multi-path beam forming can be embedded in car windows, for example, invisible to the naked eye, and enabling high broadband connectivity. Enabling connected cars in this way is, therefore, a clear use case for 5G.

The misconception with connected cars

The idea that ultra-low latency can be applied to all cars on the road is, however, a misconception. For one thing, 5G’s ultra-low latency works on the principle of prioritising some users over others. If all cars on the road required the same level of latency, it would be impossible to apply this principle.

Secondly, there’s currently no way of enforcing a rule that requires every car on the road to be 5G-enabled. In that instance, cars with 5G may not receive information from vehicles that do not have 5G. Then, of course, there are the many other obstacles including humans and animals that wander into and across the road and do not – and cannot – have 5G connectivity!

Thirdly 5G will, initially at least, be overlaid over existing 4G networks. It is worth remembering that this does mean one hundred percent connectivity as there are, notoriously, black spots on a number of 4G networks. It’s highly therefore unlikely that automated cars will rely entirely on a form of connectivity that may not be available in certain places. That’s a challenge the automotive industry must confront.

Robotics and the low-latency myth

The advent of Industry 4.0, or the industrial internet of things (IIoT), in which industrial manufacturing devices, equipment and computers are connected for greater efficiency, has seen an increase in the deployment of automation technology and robotics. As with connected cars, the ultra-low latency promised by 5G has been seen by many as an enabler for robots to receive and analyse data in real time, and act and react accordingly.

In the main, however, robots tend to have both processors and memory on board and, while low latency computation – the actual movement of a robotic arm, hand or digit – is important to a robot’s functionality, it is already achievable through fast broadband connectivity. This same broadband connectivity is also required for analysing data in the cloud, or for any software upgrades that might be necessary.

5G will provide the consistent, reliable broadband connectivity required for robotics. Its ultra-low latency capabilities, though, may not be as critical in this case as the hype from many vendors might want you to believe.

Virtualisation and scalability myths

Virtualisation will, of course, offer operators further flexibility and cost benefits. Despite what many vendors may claim, 5G technology is by no means ‘plug and play’. While it can be deployed on Commercial Off-The-Shelf (COTS) hardware, thus reducing costs, operators will still need to invest in integrating certain pieces of multivendor solutions, for a successful interoperable deployment.

Another common misconception is that, due to the high data bandwidth provided by 5G, the core network and its component functions such as Accessibility and Mobility Management (AMF), Session Management (SMF), and Policy Control (UPF), will scale up and down to a greater degree than in its 4G counterpart. While 5G core network virtualisation will certainly help operators to maintain their network, these are control plane elements, and the scaling of these components will not be affected by a higher level of data traffic, unlike with user plane functions.

5G: Are we nearly there yet?

5G is the future, and all the signs suggest that it is very nearly here. It will undoubtedly change the way that networks operate, and the way in which mobile connectivity is delivered. The opportunities it opens up for new and advanced technologies such as connected cars and automated industrial processes are certainly exciting. When it comes to anticipating its potential, however, the industry must be sure to measure what is hyperbole against what is actually real.

 

R.Ezhirpavai_PotraitPrior to joining Aricent, Ezhirpavai worked at the government aided telecommunication organization C-DoT. At Aricent, she has played a key role in development of software framework for solutions like SS7/Sigtran stacks, VoIP stacks and has contributed to IETF specs for SCTP (RFC 4960). Pavai has lad various projects and led driven product conception for clients, while working closely with engineering and business teams. As part of her previous role in Aricent Innovation team, Ezhirpavai was instrumental in enabling some of Aricent’s most successful software frameworks and solutions. Most recently, she has been involved in creating virtualized solutions and creating stateless solutions for a truly NFV based solutions at Aricent. She is also spearheading 5G solutions in Aricent.

Ericsson, Intel and Telstra deliver 5G low-latency to gamers FTW

As 5G demos move into their practical phase, Ericsson, Intel and Telstra decided to show how 5G will enable gamers to compete over mobile with no performance penalty.

A lot of professional gaming relies on quick movements and reflexes to get one step ahead of your opponent, a characteristic it shares with physical combat sports. So any technological factors that delay a gamer’s actions, even by a fraction of a second, can prove virtually fatal. The latency experienced by mobile networks has, thus far, made them impractical for use by pro gamers and this demo was designed to show that, with 5G, those dark days are behind us.

It returned data transfer latency rates of 5-6 milliseconds, four times lower than current average 4G latency speeds. The demo was conducted in Telstra’s 5G Innovation Centre in Australia’s Gold Coast, using Ericsson and Intel kit over mmWave spectrum, and professional Australian gamers The Chiefs were drafted in to provide extra authenticity

“This gaming demonstration is a real-life example of how 5G might be used in the future,” said Telstra Executive Director Network and Infrastructure Engineering, Channa Seneviratne. “Latency is the time it takes for data to be sent between two points, so it is crucial in the world of gaming when milliseconds can literally mean the difference between winning and losing. eSports demonstrates how that is possible over 5G, a benefit of the new technology that will underpin a host of use cases.”

Ericsson and Intel spokespeople said pretty much the same. While this is a nice demo it’s not clear how much demand there is for competitive gaming over mobile networks. If it’s that important then its presumably not something people will do on their phones on the way to the shops. But as an illustration of a cool new use-case that may capture the imagination of a broader audience then just those of us who obsess about this sort of thing for a living then it has some value.

Deutsche Telekom is living on the edge

The Living Edge Lab testbed, conducted by DT, Crown Castle and Altiostar, claims to be pushing the edge computing envelope.

As we get closer to 5G actually being a thing, as opposed to a flood of PowerPoint presentations and hastily arranged partnerships, the emphasis is on ‘real world’ testing, as evidenced by Huawei’s efforts in Canada. This initiative took the form of ‘an ultra-low latency mobile testbed to three sites centered on Carnegie Mellon University in the US’.

Edge computing will be a key component of 5G, especially from a low-latency perspective, as the best way to reduce the amount of time it takes a mobile signal to get from A to B is to reduce the distance it has to travel. The more stuff you can do on the edge of the network, as opposed to the core, the lower the latency, says the prevailing wisdom.

“The Living Edge Lab testbed is a major technology milestone towards use-case centric Edge Computing and will provide application developers with an early experience of the benefits of 5G technology,” said Alex Jinsung Choi, SVP Research and Technology Innovation at Deutsche Telekom. “It is a unique Edge Computing platform that leverages a fully virtualized end-to-end solution and the implementation of user-tracing beamforming antennas for the first time in a live environment.”

There you have it. They are trying out all the latest radio technologies there, are currently working with latencies down to 15ms, and seem to be focused on the 3.5 GHz band. It should be noted that Aerosmith have been calling for this kind of thing since the early 1990s and they will no doubt be gratified to see the telecoms industry finally take note.