Three privacy issues in edge computing

edge computing

The rapid development of the Internet of Things (IoT) generates massive amounts of data daily. Photographers, for example, can take 100 Mb of photos daily, while surveillance cameras can record 20 Gb of video daily.

The computational limitations of an IoT device make real-time processing and analysis of unimaginable amounts of data nearly impossible. A centralized server with sufficient computational power can process this data using cloud computing. However, cloud computing is inefficient in dealing with this large amount of real-time data due to its low bandwidth and high latency.

As a result, edge computing has emerged as a viable technology for achieving high bandwidth and low latency. Edge computing can deliver new services and applications to billions of IoT devices by offloading some of the computational power and storage to the network’s edge, such as augmented reality, video analytics, smart homes, smart hospitals, Internet Vehicles, and so on.

A cloud server with sufficient computational resources and storage space is typically located in a data center, far from most end users. Edge servers are geographically close to end devices at the network’s edge to ensure high bandwidth and low latency. Edge servers typically have more computational resources and storage space than end devices but fewer than cloud servers. The end device usually communicates with the edge server to get a quick response.

However, privacy is a critical aspect of edge computing. In edge computing, there are three types of privacy concerns.

1. Data Privacy

Most end users prefer cloud/edge storage because they can access their data remotely and easily share data; they can avoid capital expenditure on physical hardware costs; and they don’t have to worry about file and storage management issues because the cloud/edge service handles this for them.

More data is transmitted between end devices and edge servers due to the high bandwidth of edge computing. This enables the transmission of more private information. Photos, personal health records, and government data may be leaked or hacked by unauthorized users and third-party companies.

On the other hand, Edge servers are difficult to control, in contrast to cloud computing, which has a central data center and is usually strictly supervised. Although edge service providers use firewalls or virtualization to prevent data leakage, these mechanisms are ineffective in protecting users’ privacy due to untrustworthy edge storage services.

To protect the privacy of data stored in the cloud, the conventional approach in cloud computing is to encrypt users’ sensitive data before loading it into the cloud. The data can then be retrieved using a keyword or ranked keyword search. The main disadvantage of encryption methods is their high computational cost and overhead.

Furthermore, while the end device is typically connected to the nearest edge server and may be migrated from one edge server to another for a better experience, private data can easily leak during this process. In edge computing, a privacy-preserving algorithm may be run between the cloud server and the edge server or between the end device and the edge server to protect sensitive data.

2. Location Privacy

In recent years, an increasing number of applications for location-based services (LBSs) have been adopted and have been successful in many aspects, such as improving traffic, road planning, finding the nearest points of interest (POIs), and so on. Edge computing is a natural and ideal system for LBS because end devices typically connect to the nearest edge server for a better experience. Users must send queries to the LBS server to take advantage of the benefits provided by LBS. Conversely, these queries contain massive amounts of information, such as users’ locations, interests, hobbies, and so on. Untrusted LBS servers can easily access and release sensitive personal data to third parties such as advertisers.

There are two types of location-based privacy issues.

  • Restricted Space Identification: For example, revealing a user’s location may reveal the user’s real-world identity, allowing an adversary to locate the subject and cause physical harm.
  • Observation Identification: For example, if an LBS provider frequently observes the user’s bar and liquor queries, the adversary may conclude the user is an alcoholic.

Furthermore, if the user with the end device moves and the end device switches service from one edge server to another by communicating with each other, the active path information of this user may be disclosed to the curious edge server. This will make protecting our location data extremely difficult.

3. Identity Privacy

Personal Identifiable Information (PII) or user identity is information about a person that edge cloud services collect, assess, or use on demand. When users sign up for a new Edge service, they typically fill out an online form and provide sensitive personal information (e.g., name, gender, address, phone number, credit card number, and so on). This data could be stored in a centralized Identity Provider (IdP). It may later be distributed to service providers (SPs) for authorized requests, payment completion, service customization, etc. In early 2018, the Facebook data scandal resulted in the disclosure of 50 million users’ PII to a third-party company, Cambridge Analytica, for “analysis” purposes via SP. This practical example teaches us to be vigilant to protect our personal information properly. In the last decade, identity privacy issues have been closely linked to the problem of Identity Management (IDM).


We have discussed some existing privacy issues in this article, such as data privacy, location privacy, and identity privacy. There are, however, many more interesting open privacy problems in edge computing. People are increasingly concerned about protecting their personal information in this age of information overload. In the future, the privacy issue in edge computing must receive more attention.