5G networks will enable some of the most disruptive Industrial IoT applications that will shape our society and industry in a big way over the next decade; and it would not be able to do so without setting up Edge data center networks that will extend the required computational horsepower within close proximity to the machines emitting the data.
Goal: Single-digit millisecond Latency for Industrial IoT Applications
Autonomous vehicles, Robotic Manufacturing, Augment Reality and Machine Learning are some of the 5G applications that require latency to be at levels where it virtually doesn’t exist. Network latency that we have today would make it impossible to implement such applications. LTE networks are simply not ready for some of these deployment on a wide scale.
That’s where 5G comes in. 5G wireless networks will offer speeds upto 10Gbps which will allows for faster transfer of data to and from devices in the field over to edge data centers. We are talking about latency of single-digit millisecond.
Alternative to Edge Computing?
Until now the internet has been a largely a download platform for the consumers and businesses and therefore the infrastructure has been built in a manner that supports that sort of usage structure. With the advent of IoT, internet is now getting connected to sensors and actuators all over the place. Which means more data is generated and sent over the same channels built and previously mostly used for download purposes. Restructuring this whole mechanism is highly costly and probably not very technically feasible at this time. Which makes Edge Computing the only viable way of launching the world into the 5G era.
Who has the “Edge”?
Traditional edge data center providers offer some great advantages in various markets. However, there are many factors that give Telco’s an advantage over data center providers when it comes to setting up distributed Edge Data Center networks required to enable wide scale availability of 5G applications such as driverless vehicle fleets, Virtual and Mixed Reality, Robotic Manufacturing and so on.
Service Providers Advantages:
Modernized Massive Infrastructure Already Available
There exposed to data that is service providers already have pretty well distributed and spread out infrastructure assets which means they can fairly quickly turn their tower sites and regional / local offices into edge data centers while building more of them. Add small cells to that and you have added a whole new dimension to the edge network. This of course is becoming more of a byproduct of the global network virtualization and automation initiatives that service providers have taken in recent times and it gives them the head-start to be able to manage many edge sites remotely with minimum upfront costs and low OpEx.
In addition, service providers also have the interconnecting infrastructure in place to connect with the devices, other data centers in the edge network, regional facilities, as well as with the core data centers far away. In addition to providing compute and local storage, Edge data center networks leverage their backend transport infrastructure to also backhaul the data from the edge to a centralized cloud server for user analytics and reporting.
Using the remote and automated management benefits of SDN, service providers are able to manage these remote edge data center networks much more efficiently, not to mention the cost effectiveness it brings in deploying software on off-the-shelf white box appliances.
Experience with User Generated Data Sensitivities
Telco’s have stringent regulations they have had to follow over the decades with regards to how to manage the privacy of all the information that their users generate – calls, messages, numbers, fax, pictures, etc. This experience and relationship with consumers spans way back and before IoT was even conceptualized. They offer a great deal of experience and technologies to manage user generated information.
Role of other players in the Edge Data Center market
Speaking outside the realm of Industrial IoT, edge computing as an IT solution has been around for a while. The idea of moving resources to the edge is not completely new and it became inevitable especially when video streaming applications became preferable over cable tv. Edge data centers were already adopted by the large Content Delivery Networks (e.g. Google, Akamai, Netflix) who cache their content and services at the network edge through 3rd party colocation data centers and specialized edge data center service providers. For example Google has its agreements with almost all Telecom Service Providers globally and most of the Google services are cached at the incumbent network operators which saves Google hundreds of millions of dollars on the international connectivity, as each and every user do not have to reach the Google servers in USA to download the YouTube video that they want to watch.
Specialized Edge data center companies have made it possible for the video streaming application companies, as well as other time critical application providers such as Financial Services and Gaming, to reach customers in tier 2 and tier 3 markets and provide a user experience they would expect when in a tier 1 city like New York or LA where the original content servers are located. Not only it increases the user base but also saves millions of dollars of investment in terms of payments for using Telco infrastructure over long distances each time a video is downloaded.
As we enter the era of IoT where billions of gigabyte of new data will be generated, shared and stored at rates never seen before; and artificial intelligence, machine learning and mixed reality applications will proliferate at consumer as well as industrial scale. Moving resources to the edge and building edge data center networks is going to be a critical part of the network transformation efforts currently underway.
The post Edge Data Center Networks key to Industrial 5G Applications appeared first on Lanner.
This post first appeared on Software-Defined Networks, IoT And Next-Generation Infrastructure, please read the originial post: here