In the Windows Update settings, there’s an option called “Delivery Optimization.” It allows us to obtain update files from other devices, including those on the local network as well as those on other external networks, rather than solely downloading them from Microsoft’s servers.
This feature is highly beneficial for users with multiple devices.
Imagine a company with hundreds of Windows PCs. If each computer were to download updates individually from Microsoft’s servers, with each update being large in size, it would consume a significant amount of internet bandwidth. However, if the devices have “Delivery Optimization” enabled, the update downloaded by one device will be cached locally, allowing other devices to retrieve the cached update rather than downloading it again from Microsoft’s servers. Given that internal network speeds are typically fast and bandwidth is ample, this approach minimizes the impact on internet bandwidth, ensuring other devices’ internet access remains unaffected.
This feature is also immensely valuable to Microsoft, as it helps the company save substantial costs on storage, bandwidth, and CDN services.
This technology can be referred to as PCDN (Peer-to-Peer Content Delivery Network), a form of edge computing. To aid in understanding edge computing and PCDN, we’ve translated and reprinted a blog post here.
Title: What is Edge Computing?
More and more services claim to use edge computing technology. But what exactly is edge computing?
By definition, edge computing is a distributed computing architecture that brings computing, storage, and network resources closer to the data source or user, moving them from traditional centralized data centers to edge nodes in the network. This reduces the distance data travels over the network, lowers latency, improves real-time processing capabilities, and saves bandwidth.
Imagine autonomous vehicles, which require extremely fast recognition capabilities. Due to network latency, they can’t rely on sending data to a server for real-time analysis and then waiting for a response. Instead, the vehicle itself must have some computing power, exchanging only minimal data with the server when necessary. This is an example of edge computing.
Edge Computing is Everywhere
As technology advances, more and more devices are becoming connected, creating an increasingly interconnected world.
While connected vehicles are one example of IoT, the most common interactions people have are with smart homes. Smart locks, network cameras, air conditioners, water heaters, refrigerators, washing machines, dryers, and vacuum cleaners have all become common household items in recent years. More recently, smart curtains and smart lighting have also started gaining popularity.
All of these devices need to be connected to the internet. If everything were handled by servers, it would overwhelm companies and result in poor user experiences.
Take network cameras as an example. With increasing concerns about security, cameras now offer advanced features like person detection, object movement detection, and baby cry detection. These require the camera to perform local computations to make quick decisions, rather than merely recording and uploading data.
Security Concerns and Solutions in Edge Computing
From a design standpoint, edge computing offers inherent privacy advantages. Since many computations are done locally without uploading data to a server, it reduces the risk of privacy breaches during transmission or on the server. However, in practice, if edge devices lack security, they may be more vulnerable to breaches than servers maintained by professionals, leading to local data leaks.
This issue often arises with routers and cameras. Routers are used in almost every household, and cameras are particularly targeted by hackers due to privacy concerns. Most home users are not professionals and often fail to properly set up their connected devices, leading to the widespread use of simple passwords, default passwords, or no passwords at all. Additionally, many IoT devices have firmware vulnerabilities, and most users lack the awareness to keep them updated, making it easier for hackers to launch mass attacks.
Furthermore, edge devices can both compute and connect to the internet, essentially functioning as simple computers. Once an attacker controls a large number of edge devices, they can issue commands to simultaneously attack a single target, known as a DDoS attack. The rapid growth of IoT has made DDoS attacks more common and more industrialized.
Clearly, solutions must address both ends.
For manufacturers, it is crucial to use strong encryption to prevent data from being intercepted or accessed in plain text during storage and transmission. Enforcing strong identity verification and access management can prevent the use of simple, default, or no passwords. After products hit the market, continuous improvements should be made, with regular security patch updates or new firmware to prevent vulnerabilities from being exploited.
For users, attention should be paid to the security of each connected device. Proper security settings should be configured according to the product manual, using complex passwords to protect devices. When update notifications are received, devices should be updated as soon as possible to patch vulnerabilities and reduce the risk of attacks.
Is PCDN Considered Edge Computing?
Videos contain more information than text or images and occupy more bandwidth. The rise of traditional and short video platforms has driven up network bandwidth demand. Distributing and caching videos on terminal devices, so that other users can load resources from the nearest terminal when accessing the same content, has become an effective way for platforms to reduce costs and increase efficiency.
PCDN (Peer-to-Peer Content Delivery Network) can be considered a form of edge computing.
Suppose neighbors A and B are both users of a certain video app and have both recently become obsessed with the same video series. After A finishes watching the video, the video files are not immediately deleted but are cached in the app. When B clicks on the same video, the server does not directly send the video file to B’s phone but instructs B’s app to retrieve the resources from A’s app. If the videos B needs are cached on A’s phone, they can be retrieved from A’s app.
Originally, PCDN was a win-win design—the platform saves computing, storage, and bandwidth costs, while users help each other load resources faster. However, specialized companies and organizations have gradually started exploiting PCDN for profit, leading to its misuse.
One scenario involves home network TVs or mobile apps continuously uploading data in the background, saturating the upload bandwidth of home internet and causing network congestion. Network TVs, in particular, can cache resources while playing videos and often use Android’s screen-off standby mode instead of actually shutting down, allowing continuous data uploads that benefit TV manufacturers and app developers.
Another scenario involves video platforms outsourcing resources, with third-party companies attracting ordinary users to use their home bandwidth for data uploads and distribution by promising commissions or subsidies based on the amount of data shared. Many “money-making routers” and “money-making boxes” on the market are such hardware products.
ISPs have started cracking down on PCDN, identifying its use by analyzing discrepancies between upload and download traffic and issuing warnings or bans. In response, some people profiting from PCDN have begun downloading internet resources to balance upload and download traffic, evading ISP detection. They often download resources from CDNs due to the speed and fewer restrictions. For more information, refer to the articles previously published on this site:
- CDN Traffic Manipulation: Making My Financial Struggles Worse
- Unbelievable: Shanxi Unicom IPs Continue Stealing CDN Traffic
Many technologies are designed with good intentions, but when someone discovers they can be profitable, they are often abused, harming legitimate users. This is also a problem that needs to be considered for the future development of edge computing.