Hyperscale data centres have dominated the market since the introduction of cloud services but do not offer the best solution for latency sensitive applications.
This has given rise to edge computing, a market which is accelerating rapidly as many new services now demand the lowest latency possible and cannot be supported by the limited number of hyperscale data centres – located predominantly in the South East of the UK. With this, Gartner estimates that 75% of data will be processed at the edge by 2025, outside of traditional centralised data centres.
Edge computing provides distributed computing and storage resources closer to the location where they are needed and targets new business opportunities that provide support for specific application use cases.
However, it is important to note that edge computing in edge data centres complements rather than competes against public cloud services. Therefore, developers reliant on the lowest latency possible must consider the best place to deploy and support new services and need to rethink the network architecture.
These new services include:
- Cloud Gaming – Nearly all online gamers have experienced ‘lag’ when gaming, resulting in a high amount of player frustration. Latency is critical for gamers, so the lower the latency the better the experience. High latency can disrupt (or even stop) a game in mid-stream while low latency provides smooth, buffer-free gaming. Jitter can also cause problems for gamers if it’s too high. Jitter occurs when latency is inconsistent rather than stable and its effects are similar to high latency. High jitter can lead to choppy or laggy gameplay and/or distorted audio and chat functionality, while low jitter can keep games running smoothly.
Packet loss is another consideration and occurs when packets of game data are lost in transit to and from the user and cloud service. Packets may also arrive out of order which can lead to a game that’s unintelligible. Like jitter, packet loss can lead to an extremely frustrating user experience and can make the difference for a gamer between mission completion and mission failure.
The deployment of Cloud Gaming at the edge will significantly improve latency, jitter and packet loss. The Cloud Gaming platform that can deliver the best player experience will win out over the competition.
- Augmented Reality and Virtual Reality – Both these services need low latency to deliver the best user experience. Low latency is critical when using a head mounted display as the human brain will detect any delay. This can also cause simulator sickness and virtual content to be misaligned with real life features. Known as motion-to photon latency or end-to-end latency, the higher the latency the more unrealistic the VR world seems.
- Automated Intelligence and Machine Learning – As well as demanding powerful computing and large amounts of storage, AI and ML applications also require low latency services. Latency impacts performance, while data intensive applications with multiple storage architectures require low latency as well as high throughput. These applications include compute-intensive machine learning and dense neural networks. For AI and ML to work well they need to be able to respond as quickly as a human does. For example, in call centres AI decisions are supposed to help human operators achieve better outcomes, however any delay causes frustration for the operator and the customer they are trying to help. By deploying these services in an Edge data centre instead of backhauling though several hubs to a hyperscale data centre will improve performance and reduce network congestion. Even small network delays, in the order of tens of microseconds, can impact application performance significantly.
- IoT – this widely spans a huge range of solutions and a lagging connection can have disastrous consequences. In the same way as humans find a slow website very irritating, in the world of IoT a delayed response can also be very problematical. Any delay in sending an instruction from an application situated in a data centre to the IoT device to carry out a task will cause problems. Highly dynamic IoT systems require faster control and short round-trip times (RTT). For example, in Industrial control which includes a very broad set of applications used in various industrial verticals, any delayed messages are typically considered as lost.
- Automated Vehicles and 5G – Edge data centres are playing a significant role in the development of the infrastructure needed to support Automated Vehicles (AV). These will not work in isolation to other road users and not all processing will take place on the vehicle. Rather, 5G will allow the AV to connect to Edge computing resources for such things as updates on local traffic conditions and parking zones while the uploading of 4K video for storage will soon become standard practice. Services such as these will work in conjunction with Smart Cities deployments as they are rolled out by the Public Sector. With 5G latency now comparable to fibre, the closer the application to the user the better and safer the experience.
- Healthcare – The new generation of ehealth applications offers tremendous advantages for offering clinical expertise to a distributed population. Virtual consultancy has the potential to reduce travel, allow fewer visits to clinics and ensure early diagnosis. However, the success of these services can be directly impacted by network limitations causing high latency and jitter due to the backhauling of data traffic 000’s of miles to centralised cloud services.
Tele-Surgery using robotic technology has inspired several treatment models that are bringing high-level care to underserved patient populations and expediting time to treatment. One popular model being adopted is to connect a network of hospitals to one centralised location in a hub-and-spoke model. Here, an edge data centre can securely support ehealth applications with the lowest latency possible. For example, physicians at a health system’s flagship location can perform remote procedures via tele-robotics on patients at ‘spoke’ locations, dramatically increasing the volume of patients that one physician can treat, allow patients to stay close to home while undergoing treatment, and enhance patient outcomes for acute conditions through more rapid access to care.
- Video Conferencing – This has become a very popular application and is used daily by millions of people. However, high latency on video calls causes significant level of frustration with time delays often resulting in participants talking over one another. Locating servers closer to users in edge data centres will improve quality of service and alleviate such time delays.
Find out more about our solutions for Content Delivery Network providers, for Gaming and for Industry by clicking here.