There and back again – After the Cloud, the Fog

Everything old is new again. That applies to most industries, trends and businesses, so why wouldn’t it apply to how we use resources and where they are placed.

A history lesson

In the 1970’s, IBM developed the first time sharing service implementation via virtual machines and the VM OS.

A few years back, everyone was building data centers.

Then, computing power and data storage were moved to a place everyone called “Cloud” but no one actually knew what it was and that in fact represented a new name for an old dream Douglas Parkhill was writing about it in 1966 in “The Challenge of Computer Utility”. The term became popular starting 2006, when Amazon launched its EC2 (Elastic Cloud Compute) service. In 2008, Microsoft followed the footsteps and launched Azure, their own Cloud service and in 2013 IBM announced the acquisition of SoftLayer, forming the IBM Cloud Services Division.

cloud floppy father

IoT is the new hit

Now, there’s the mighty Internet of Things, which promises to connect everything, but brings us back at least to the partial decentralization of resources and leads the way for the so called “Fog”. IoT is estimated to connect approximately 50 billion devices by 2020, according to IOT Analytics.

IOT Analytics device forecast
IOT Analytics device forecast

These devices are to generate approximately 2 Exabytes of data (2 quintillion bytes or 2 billions of billions of bytes) on a per daily basis.

What is Fog Computing?

The term is actually a rebranding of Edge Computing or Grid Computing, which means having a virtually distributed supercomputer (made up of small computers connected to a network). This is mostly used for computationally intensive scientific, mathematical, and academic applications, through volunteer computing, and it is used in commercial enterprises for such diverse applications as drug discovery, economic forecasting, seismic analysis, etc. So in the Fog, there are lower end devices, that dispose of fewer computation power and resources, but that accomplish the following key tasks:

  • get the most time sensitive data analyzed close to WHERE and WHEN it is generated instead of sending vast amounts of data to the cloud.
  • prioritize types of data based on policies and such serves applications according to their needs (EQoS – Edge QoS)
  • send only selected data (that has been pre-processed) to the cloud for big data analysis and longer term storage

Shortly, Fog Computing “extends Cloud computing and services to the edge of the network”, creating a sort of distributed Cloud.

Why Fog Computing?

The answer is quite simple: by the time the huge amount of data generated by all the connected devices makes its way to the cloud for analysis, new data is generated and the opportunity to act on what you have might be gone. This doesn’t mean that Cloud computing is gone. Not at all, the two complement each other in a way that makes data analysis and information extraction as fast and easy as possible.

Moreover, there are the added benefits of security and protection of networks from attackers, as well as the reduction of operating costs through automated, policy-based provisioning, visibility, and management.

IoT + Cloud = Fog?

Something like that, yes. IoT gets “things” connected, the Cloud analyzes and stores big data and the Fog helps these two get along. Now, how do you build it? Well, the first to come up with a solution was Cisco, with the Cisco IOX, combining a PaaS (Platform as a Service) with an IaaS (Infrastructure as a Service) and on top of that using SaaS (Software as a Service). Madness, but let’s see how it works:

IaaS in the classical sense provides storage and computing resources in the cloud. When it comes to the Fog, it is used to provide a deployment space for IoT apps on Fog nodes. The fog nodes may in this case be routers/switches/access points, etc.

Cisco IOX actually stands for the combination between IOS and Linux, and supports APIs (Application Program Interface) that can talk to IoT devices using any protocol and can send data to the Cloud by translating the proprietary protocols into IP. The development of new apps is facilitated using this platform, as well as their mass deployment of features. For example you have a fitness watch company, you give the watch for a small price but limited functionality and if the buyer wants full features, you can make simple software changes remotely, actually offering MaaS (Metal as a Service).

Cisco IOX
Cisco IOX

AWS IoT and no Fog?

Amazon announced its IoT on the 8th of October last year, with the buzz line “Cloud Services for Connected Devices”.

“AWS IoT is a managed cloud platform that lets connected devices easily and securely interact with cloud applications and other devices. AWS IoT can support billions of devices and trillions of messages, and can process and route those messages to AWS endpoints and to other devices reliably and securely. With AWS IoT, your applications can keep track of and communicate with all your devices, all the time, even when they aren’t connected.” – AWS IOT

AWS IoT makes use of existing services, such as Lambda, Kinesis, S3, Machine Learning, and DynamoDB to build IoT applications that gather, process, analyze and act on data generated by connected devices. The main advantage in comparison to the Cisco Solution would be that there’s no infrastructure to manage.

Nevertheless, AWS IoT is more focused on the consumer that has non-industrial needs, while Cisco would be more focused on the large scale industrial market.

One final point

Fog Computing doesn’t only work for enabling IoT. Anything that needs dynamic bandwidth allocation or limitation, content delivery or filtering, tracking user data and action, can be done in a decentralized manner, directly on the network devices at the edge of the cloud.

For example, as you walk into a shopping mall let’s say, and you connect to the Wi-Fi, you can start receiving messages about promotions, discount codes, your traffic patterns can start to be tracked, the data you generate can be transformed into useful info for the shops you enter, and so on, thus enabling content delivery as an integrated service. And all this can be accomplished just by by implementing decision algorithms and apps at the AP/ Gateways level. Goodbye privacy!

References and further reading:

Facebook Comments

Leave a Reply