‘Welcome To The Edge!!!’–What Is Edge Computing???

[Note: This is an evolving, continuously updating post for my research purposes to learn more about SASE, or Secure Access Service Edge. First, I’m laying the groundwork for what ‘edge computing’ is.]

Edge Computing

The term ‘edge’ is all the rage these days.

From a infrastructure-centric point of view, “Edge computing is at it’s essence cloud principles applied at the network edge close to the user“[3]. It can include:

Virtualization (Compute virtualization, Storage virtualization, Networking virtualization)
Resources On Demand
API Driven Approach
Automated LCM Life-Cycle management
Use of Commodity hardware [3]

These are some of the powerful core cloud basic principles that make the network edge highly flexible and programmable.

Edge can be considered the convergence of IT and telecom (including mobile) networking. It allows network operators to open up their networks to new opportunities and value chains.

NOTE! Though we a now talking this mash-up convergence of IT and mobile telecom, it is important to note that this is NOT to be confused with the mobile network edge technology. Remember EDGE cell phone technology from the 2000’s???

EDGE
The letter E represents the Enhanced Data rates for GSM Evolution (or EDGE) network. The network started to spread in popularity sometime in 2003 by offering speeds that were almost three times faster than any of its predecessors.
It supports a maximum speed of 217 kilobits per second, so even though it’s significantly faster than G network speeds, you’ll still struggle to browse a modern website or watch YouTube videos in anything but the lowest resolutions.
That said, there are now 604 EDGE networks in 213 countries, making it one of the most widely-used mobile internet technologies in the world. It was the final widely-used network before 3G came into prominence, so it’s often referred to as 2.75G. [6]

Edge computing is “a distributed computing paradigm that brings computation and data storage closer to the location where it is needed to improve response times and save bandwidth. It is a topology rather than a technology”.[1]

So, look, a quick primer and/or recap…But BASICALLY when you have only one computer, everything (including all resources, computational power, storage) is stored on only that one machine, and everything is great! And since computers are so great, the only thing better than one computer, is multiple computers!!! And when we have multiple computers, we need a way for them to communicate. This has traditionally been the client/server relationship, whereby individual end-users, the clients, requests the services from another computer, the server. Cloud extends this client/server relationship across greater distances, helping to remove that distance barrier. So now computers could network (connect with) each other over greater distances than just a physical office, for example. Now we could network computers across the globe! But with increased distance also come issues like increased lag and many other issues that come with sending data. NOTE…remember that although this data may seem intangible to us, just bits of 1&0’s essentially, it still adheres to principles of moving and sending data whether through wired cable from a telegram to telephone to wired internet to wireless internet, wireless mobile and even satellite. SO, now…we are in the midst of a transition, or one could even think about it as an advancement or progression of the cloud paradigm to bring us to edge networking. Edge networking now thats those resources from the cloud, or the servers “way out there!” and brings some of those resources back closer to the user where it’s more beneficial for all-around. It kinds of optimizes the power of everything surrounding to work together for a synergistic network effect.

As this edge computing infrastructure grows with more consumers jumping on to 5G networks along with an unprecedented rise in IoT (Internet of Things) devices and “Jetson-age” autonomous technologies, including cars, this growth will fuel it’s own network effect. Expect to see similar life-altering applications, much like how social media apps were created and fueled by the advent of 4G, but expect his shift to be MUCH MORE DRAMATIC.

So, edge doesn’t compete with cloud in a sense of, “What’s better!? Cloud computing or edge computing???”, but rather each complements the other, and allow each other to shine, leaving the desired result or application to determine which approach is better for a certain situation.

Gartner, Inc. submits the following scenario for their vision of an Edge future:

“…Consider the ever-increasing compute power and richness of application services offered by the hyperscale cloud. Then, multiply it by millions or even billions of remote nodes, many with their own compute power and the ability to interact with the physical world around them. Imagine these distributed nodes and their contect-specific interactions achieved through sensors, combined with distributed machine learning and analytics, learning, processing and reacting locally. At the same time, they’re taking advantage of the core to further analyze the data and activities of the multitude of connected devices as a whole.
Is this purely aspirational, a wild idea from a tech researcher, or science fiction? None of the above. A teenager with a credit card can begin building such cloud-integrated edge solutions today. So enterprises certainly can as well. Although the patterns are nascent, and the technology is perhaps not yet industrial strength, they will be. Soon. Welcome to the conjoining of the edge and the cloud. [2]

Quite a dream being sold!

And

Multi-Access Edge Computing (MEC) Special Report

Some points:

Multi-access edge computing(MEC) is the concept of bringing compute, storage and analytics capabilities closer to the end user, whether human or machine, and create a service-oriented cloud environment at the network edge that can be monetized with new application types. [4]

MEC is being explored in 4G but is a foundational technology built into 5G; it will likely be pushed outward in phases and could men either a more aggregation-point-heavy architecture for mobile networks or could eventually push compute all the way to the tower. [4]

Network operators are trialing MEC in a variety of ways, and some of them are working on the network transformation (including re-designing their central offices to serve as data centers, implementing virtualization, cloud-native re-architecting of applications) that will be foundational to MEC. [4]

Challenges for MEC deployment include physical aspects such as siting, power, cooling and security of MEC locations, but the greater hurdle may be in establishing mutually beneficial relationships with application developers and institute seamless application deployment processes. [4]

How to think about edge?

Edge Computing Use Cases

3rd Party Edge Applications
- Real Time
  - (AR/VR, Connected Cars)
- Cost Reduction
  - (Video surveillance)
- Immersive
  - (4k Video, 360 video)
  - Need extremely high cost throughput
  - This could be the content needed for tools like Facebook’s Oculus.
- Self-Contained
  - (Stadiums, Concerts, Airports)
  - Network slicing place a critical role here.
- Private/Enterprise
  - (RAN-Radio Access Network to Enterprise breakout)
Operator Applications
- Analytics
  - (Data Reduction)
  - Collecting large amounts of data but sending, only using only the necessary data required for that specific instance, or what I’ve come to think of as each ‘unique exchange’.
- Compliance
  - (Copyright, Geo-placement)
- Security
  - (Securing the core)
  - Moving security close to the source
- NFV
  - (Access, Security, Service enhancement VNFs)

What is it about edge computing that checks the boxes for all this use cases?

Benefits of Edge Computing

Ultra Low Latency
- With Edge computing, typical latency (think lag) is in milliseconds, significantly lower compare to centralized DC.
- Low Latency, or low lag, the time it takes for a signal to be transmitted (sent and received) back and forth between the client and server, or now in today’s connected world, think the back and forth of data exchange needed between a “requesting computer” and the “providing computer”. ONLY, in today’s increasingly connected environments, the always-on, continuous flow of data, typically needing to be constantly updated because of the nature of today’s computing devices. This will increase with the increasing rollout of 5G combined with the RAPID increase of IOT devices. Its the dawn of a new era, and one that we are already behind on security-wise.
- is very
High Network Throughput
- Content is either locally generated or cached
- Network throughput can be orders of magnitude greater than from core DC
Context Awareness
- Edge has access to radio network
- Information provided by the RAN can be used by edge apps
Data Filtering, Reduction
- Data analytics apps at the edge can substantially reduce the amount of data sent upstream.
Compliance
- Edge applications can help with privacy laws and data location laws.
Security
- CSPs can protect their networks against attacks from user equipment (UE) or customer premise equipment (CPE) using edge applications.