From clouds to fog

Metaratings
03 Jun 2014
00:00
Article

Cloud, big data and SaaS have now been the hyped key words in ICT services for a few years. In a certain way they mean centralized architecture, i.e. most computing power, data and services are collected to a few central places. And then different people and organizations can use those services in a browser or mobile. Now we start to see more products and discussion about alternative solutions, like fog. Is it time to see more decentralized solutions and can it also change the power balance from existing giants like Google, Amazon, and Microsoft?

Is it really so that it makes sense to centralize all data and processing to a few places? It requires very fast and high capacity Internet connections, it doesn’t utilize computing power of different devices (that typically have quite a lot processing power) and problems in a cloud or in connections to the clouds basically paralyze all services. And we are just in the beginning of the data collection, processing and communications era. Internet of Things, IoT, will bring more devices and data to the internet.

This is not a new discussion topic. I remember from the late 1990’s a strategy paper by Ericsson, that tried to clarify pros and cons of two main alternatives: to have all power, intelligence and data in central systems, or distribute it to devices that are connected to the network. We can also say that mobile apps are a kind of example of the more decentralized model. They hold much more local processing and data than Internet browsing. One reason for the success of apps has been the usability and how it is also linked to reliability, availability and speed of data connections.

If we imagine the situation in the next 10 years, most of devices at home have a small computer, can collect data and make actions based on data. Is it really necessary that each sensor and device sends their data to a center maybe thousands of miles away? Or could we have some devices at home that collect the main data and make decisions and distribute them to other devices. The same is valid with our car with hundreds of sensors and decision-making units. And our car and home are simple cases, if we compare to a larger factory, supermarket, airplane or public transportation control system. Is it just too much data to transfer, too expensive to build the infrastructure and too prone to technical problems and security risks?

Especially Cisco and IBM have now been active with the ‘fog concept’, i.e. distribute data and processing to local hubs. For Cisco it can mean smart routers that can take care of data storing and processing on the local level. I think fog is not yet an established term, but I would say it is an umbrella term for architectures that distribute data, intelligence and processing power to several layers from the edge of the network to the central cloud services and try to optimize the use of processing capacity and data transfer resources. These companies also see that Internet of Things can change the balance of the ICT business (I wrote about this earlier, read here) and new concepts are preparation work for the IoT future.

Privacy has been a hot topic, and it is definitely linked to the centralized cloud architecture too. It can be relatively easy for ‘big brother’ to gain access to and control a few clouds, but it is more complex monitor millions or billions of small devices around the world. It is an illusion to think that a fog model would remove privacy concerns, but it would definitely change the models for monitoring, but also protecting data.

Can this change the business power in the industry that is now dominated by the likes of Google, Amazon and Microsoft? I think Cisco and IBM hope so. This could also present an opportunity to carriers, when they manage these local level solutions and devices. They could offer and host the local fog solutions. But carriers are typically excellent in missing opportunities. We must remember that Google already now offers some internet connections, plan their drone based connections and self-driven cars. Facebook and Microsoft are also active with new solutions. It is easy to believe that all data and processing cannot go to a few clouds and we’ll see also more smart decentralized solutions. But, whether or not it is a disruption in business and power balance between companies, is hard to say.

Related content

Comments
No Comments Yet! Be the first to share what you think!