In my last blog post, part one of a two-part post, I interviewed CTO Chris Rivera for an overview of cloud, fog and edge computing. In this second part of our conversation, I press Chris for more detail about different use cases for these technologies along with how some of these challenges can be addressed.
Mary Clark: So, Chris, to recap what we established in our first conversation . . .
- In cloud computing, the storage and processing power is in the cloud, which supports the devices.
- In edge networks, the devices closer to the action process the data but feed information back to the cloud for further analysis.
- And in fog networks, the main cloud supports a network of mini-clouds that are closer to the end user, where they can undertake processing and some storage with lower latency than the main cloud, which they all serve.
Chris Rivera: Correct. And the choice of the right network setup will vary between applications and industries.
While there is no one-size-fits-all approach, there is a lot in common across different use cases. To determine the best approach, it’s important to ask the following questions before proceeding with network design:
- How much communication latency is acceptable for your application?
- How much processing power do you need at a device level?
- How many partners need access to your data or your network?
- How do you control access by these partners?
“Given the sensitivity and importance of the data that could be involved . . . companies will need these networks to be always available, always bandwidth-capable, and always secure. Without question, the public internet can’t meet these demands.”
Mary: Clearly, an extension of these questions is to determine the communications network that will underpin the chosen infrastructure.
Chris: Absolutely. All of these networks rely on two-way communications – sometimes local, but often national or global IP connections. Given the sensitivity and importance of the data that could be involved – whether business-critical operational data, safety-related information or financial data – companies will need these networks to be always available, always bandwidth-capable, and always secure.
Without question, the public internet can’t meet these demands.
Mary: Put this into context for me. Can you give me an example use case?
Chris: Sure. In the utilities market, equipment sensors might be monitoring and controlling units within a plant but reporting back on their conditions to a private or hosted cloud. In this case, it will be absolutely vital that the links between the sensors and units are always and on and always secure, as these links are vital to feeding information back to the devices that determine the way they manage the equipment.
Just recently, there were reports of an incident involving the national power infrastructure in Ukraine, where the attackers got inside the network to manipulate the controls and cause widespread power cuts and failures. There was a similar case in the U.S. last year as well. For networks like that, the attack surface available to hackers is simply too wide to defend to be allowed to rely on public internet connections.
Mary: So, what’s the solution? Do these networks need to be private and closed to outside groups?
Chris: That’s partly true. The networks do need to be private, but in many cases they will also need to be open.
Let me explain by looking at the fintech market. The PSD2 regulations coming into play in Europe will require financial institutions to open up their networks to third-party vendors. Lots of small, innovative companies will be looking to access these networks to sell new consumer services or provide innovation to a banking partner.
Obviously, these networks that process data in the cloud for millions of financial transactions every moment need to be secure and protect consumer privacy. And opening them to partner companies will need to be tightly controlled.
The host companies will need certainty to know, at all times, exactly who is connecting and what they are allowed to see and do. Again, that privacy cannot be guaranteed if those connections are coming in over the public internet, with its vulnerability to attack over such a wide surface.
Business-critical networks – whether cloud, edge or fog – will need to be connected using communication links where the access is restricted, so the identity and rights of the people, applications and devices accessing them is strictly controlled. These networks must be protected from outside interference, with guaranteed availability and sufficient speed and capacity to deliver against well-defined service-level agreements.
Those are the networks we expect to see rise and dominate going forward.
Mary Clark is a former Chief Corporate Relations Officer and Chief of Staff at Syniverse.