Internet data privacy tug of war

Frequent outbreaks of data leaks

Due to frequent data privacy incidents, individuals and the government are paying more attention to data privacy. According to the news at the press conference of this year’s TWO sessions, the Standing Committee of the National People’s Congress has included the enactment of personal information Protection Law in the current legislative plan. Before the GDPR, and after the new law on the road, in this era, as individuals, what expectations should we have for our data privacy? As enterprises, how can we ensure that our apps can effectively protect users’ data?

Data leakage caused by system security

In the second decade of the 21st century, with the development of technology, our data storage and processing capabilities have made great strides, and the consequent data leakage problem is also becoming more and more serious. The following table shows the 10 major Data leaks reported by mainstream foreign media from August 2018 (Data Dispose 101) :

It can be seen that even the rich and technical strength of the big factory, their API exposure to the standard, do countless penetration testing & defense, but also can not avoid “data leakage”. It is also because it is a large company that has a large amount of user data, data leakage will be “extraordinary” once it happens.

But, according to???

To understand why, start with the hackers’ methods:

Hacker’s methods of attack

To borrow a diagram from Trend Micro, hacking generally falls into four stages:

I research

From the network, system, people and other angles to find the target vulnerabilities, this aspect of defense because of the involvement of too wide, most organizations are based on education and self-examination;

II attack

There are roughly two ways to do this

1. “Front door assault” : Attacks on infrastructure that traditional firewalls mainly defend against.
2. Rogue: Refers to “social engineering” attacks, such as email phishing, pirated software trojans, fake WiFi phishing, social networking impersonation, and even getting in the company in the flesh. . If it’s a worthy goal, use a lot of valuable 0Day bugs:

• Stuxnet APT targeting Iran’s nuclear facilities.
• Poison Cloud Vine (APT-C-01) : An 11-year attack on China’s defense establishment.
• Blue treasure mushroom (APT-C-12) : An 8-year attack on China’s nuclear industry and scientific institutions.

III drag library

Once you break into a company’s office network or even a server, the address and password of the database are usually put in the code in plain text due to the history of the database (more on this later). Even if you don’t find the password, just copy the data file. All that remains is to bridge the way back to the valuable data. Because of its simple development and strong adaptability, most of the businesses involving user privacy use DBMS (Database Management System, relational Database) as storage. So, whether “front door assault” or “crooked” sword DBMS.

IV and hide

1. Only a few data breaches are exposed, or long after the attacks have taken place, the data has been hacked layer by layer until its value is exhausted.
2. For more valuable targets, hackers will choose to lurk, constantly obtain more valuable data;
3. As most of the Intrusion Detection System (IDS) itself also uses simple file storage or ordinary database……

Data leakage caused by management negligence

The most influential event was the “Cambridge Analytica data breach” of Facebook in March 2018. Cambridge Analytica collected data from 87 million users without Facebook’s knowledge. Some analysts point out that Trump’s victory in the 2016 U.S. presidential election may have a lot of connections with the leaks, and behind the leaks there is the shadow of collusion with Russia. The scandal sent Facebook down 7% that day, wiping out more than $36 billion in market value.

What do service providers do

System security protection has the following principles:

The best entry point for security is in the entire system
Point of architectural convergence

First line of defense

It is natural that most of the security efforts are focused on the API Gateway.

The current App development includes the development of Web App, Android, iOS, Applet and other terminals. For the sake of development efficiency and convenient management, the back end generally provides unified API for all the above clients to call. Due to the difference in the technology stack used by each developer, the developed API also has a variety of flowers. This has also led to various apis without adequate security considerations becoming the most vulnerable areas for data breaches.

Most of the common defense mechanisms at the API Gateway level (including the “XXX website guard” that I once participated in the development of) are limited in effectiveness due to the complexity of the protocol layer and coding layer.

Second line of defense

If we rethink the basic architecture of back-end development today, we can see that another convergence point in the system is the database.

Almost all of the existing viable DBMSS are products of the last century. Due to historical limitations, there is a lack of statement-level fine-grained permission control. In addition, due to the theoretical and engineering difficulties of the database itself, for a long time, we have been adding middleware to the periphery by means of DB Proxy, for example:

• ProxySQL: a solution promoted by Percona, a well-known vendor in the MySQL database field;
• Vitess: YouTube has been working on this project since 2011. The engineering quality is very high. Most of the SQL parsing libraries for Go today have roots in this project;

But even big players like Percona and Google have failed to make DB Proxy a mainstream solution. The reasons are as follows:

• DB Proxy scheme has a negative impact on system performance and stability because many execution steps and databases are repeated.
• The original intention of DB Proxy is to solve the problem of database horizontal expansion, which solves part of the problem, but introduces more problems.

Line of defense

As the “iron wall” was built, the porter’s key became hotter.

Zuckerburg admits that Facebook’s team was aware of the mass collection of user data before Cambridge Analytica broke, but didn’t take it seriously enough. In his opening remarks to the European Parliament hearing on Cambridge Analytica, Zuckerburg acknowledged again:

“What is clear is that over the past few years, we have not done a good job of preventing our tools from being used to do bad things. Whether it’s fake news, foreign interference in a presidential election, or misuse of user data by developers, we’re failing to meet our responsibilities. It was clearly wrong and I apologize.”

But to be fair, Facebook is an industry leader in terms of both awareness and technology when it comes to protecting user privacy. But still did not avoid the tragedy, which also makes the data security issue from the Internet practitioners, into a hot social issue.

Everyone’s data rights

Overbearing user terms

For a long time, the whole society’s disregard for data rights has led to the birth of many shameless user terms. The following are the user terms of an App in 2017, roughly meaning:

1. All rights to your videos and comments are mine;
2. If you get me into trouble for any reason, you’re going to pay me back;
3. Dear users, to register this App is to admit that I am your father, and only father.

Screenshots:

Images from:
www.getit01.com/p2018011516…

Data ownership

As the “new oil” of this era, data, especially centrally stored data, is fatally attractive to hackers. As an old Chinese saying goes, “If you are not afraid of being stolen, you are afraid of being remembered by thieves”. The more fundamental question is:

In the early days of the Internet, very few people realized the value of data, and almost all vendors induced users to click, intentionally or not
Agree user agreementHijacking our data, and the trend is toward more centralization.

However, linguistic and political divisions have prevented the continent from producing an influential Internet oligarchy during this Internet wave. This, in turn, makes the EU’s thinking about Internet privacy more on the side of ordinary users. It seems inevitable that laws such as GDPR, which are so draconian on Internet firms, were born in the European Union.

GDPR requirements and limitations

In brief, GDPR has the following three requirements for Internet manufacturers:

1. Users should have the right to know how their data is being used;
2. Users have the right to be forgotten and the right to request the deletion of all their data on the site;
3. Violations could result in a penalty of 4 percent of the previous year’s global annual turnover, or at least 20 million euros.

GDPR is generally a directive, with no detailed guidance on how and to what extent vendors should do it.

For example, how far should right-to-know vendors enforce right-to-know? Since the data is stored in the database controlled by the Internet manufacturer, how to detect and collect evidence of user data abuse?

For example, the right to be forgotten is actually a very difficult feature to implement, as Zuckerberg mentioned at a Harvard University symposium in February 2019:

Zuckerberg said tools to enable Facebook users to wipe browsing history data from Facebook are still in development. And he acknowledges that developing such tools is complex and will take a while.


“Channels of information go deep into all kinds of different systems,” Zuckerberg explained. The idea here is that Facebook collects so much data from so many places that erasing the record of it all has become much harder than Facebook originally thought.

Status quo and practice of Privacy Ordinance in Hong Kong

In terms of laws and regulations, the author thinks the Privacy Ordinance of Hong Kong, China is a good one. The core principle is: the purpose of data use and collection is consistent; Delete immediately after use. GDPR of the European Union is similar, and even more operational, which is worth our reference.

As a window of China’s reform and opening up, Hong Kong is also very advanced in the exploration and practice of citizens’ data rights, and people from all walks of life pay higher attention to this aspect than the mainland. As early as 2013, a software named “Qidiyou” was launched in the Hong Kong mobile application market. Similar to the current “Tianyan Zha” and “Qicha Zha” in the Mainland, it is an App that can query industrial and commercial registration and litigation information by collecting data open to the public.

The developers of an App that used Open Data in good faith were outraged when it was removed from various App markets due to various privacy violations complaints.

With the progress of technology, when Open Data, which has not been widely paid attention to before, is suddenly presented to us, it will surprise the public. But as a technologist, I still believe that the progress of technology will eventually make everyone’s data rights better protected.

The future of technology empowerment

Artificial intelligence (ai)

Now widely used AI (Artificial Intelligence) can be roughly understood as known functionsMulti-group results of

.

I’m going to do everything I can to get an approximationAs far as possible, letAnd statisticallyAs close as possible. This technique is now widely used in Internet commerce to recommend various products to you, which is an equation like this:

​

Then put the item with the highest probability score in the best position, so that your purchase behavior can be “predicted” or even “controlled” to some extent.

via
capgemini.com/blog/

Based on the same principle, Cambridge Analytica collected the social network data of 87 million users to analyze each person’s likes and dislikes for various labels. To take an inapt but understandable example: Suppose Cambridge Analytica was entrusted with boosting likes for Trump-related stories. The strategy was to target everyone with “targeted news” by analyzing that you were a person who cared specifically about affirmative action. Then find out some news conforming to “Trump” and “support for equal rights” from the news database, intentionally or unintentionally show it several times in your social media Timeline, and “Intercept” your consciousness just like in the movie Inception, so as to have a certain statistical impact on the approval rate.

Block chain

The birth and widespread attention of blockchain technology has brought a silver lining to the problem of data ownership

Users regain control of their data

To take a simple example: in the future, our personal data will be stored in a decentralized cloud database, like Bitcoin, where we can completely control our data with a key. A very feasible data usage specification can be formulated in conjunction with it, the core of which is to require manufacturers to strictly implement the principle of restricted use and deletion after use for users’ data. For example, assuming Facebook is our trusted vendor and follows the rules, we can give Facebook a key that we authorize to read only our name, age, and friends list. At the same time, every reading of our data by Facebook is recorded, and if we find Facebook doing something we don’t want with our data, we can revoke the key and hold it accountable. Since the user is in full control of his private data, the user’s database public key becomes the true “Internet ID.”

The data flow

In the digital age, many people say that data is the “new oil” of this era, but unlike oil, data can only release the huge value contained in it when it is circulated. At present, there are still a large amount of data sleeping in the hands of the government, enterprises and individuals, the degree of data sharing, opening, circulation and application is far from enough, and there is still a huge gap between maximizing the value of data to be shared by the whole society. The biggest headache for startups or research institutions in the big data industry is the lack of data. Ordinary users can license data to these research institutions in the form of fees. This creates a win-win situation and avoids the data hegemony of the giants that store so much of our data.

conclusion

The issue of Internet data rights is a problem that needs to be solved by social cognition, regulations and technology and promoted by the whole society. The Internet is a little broken in this regard, and we believe it should be more than that.

The authors introduce

• Auxten: Worked for Baidu, 360, Tongcheng Yilong and The Fourth Paradigm. Little knowledge of server development, back-end security, Android infrastructure, infrastructure, AI. CovenantSQL: a decentralized database dedicated to improving users’ Internet data rights.
• Lao Si Fu: A young teacher of a university in HK, who has been paying attention to big data and block chain for a long time and exploring the application and teaching of media.

Refs.

• Zhang Yesui: The enactment of the Personal Information Protection Law has been included in this year’s legislative plan for an early introduction
• The biggest data breach in Facebook’s history
• The speaker of “You” complained
• HK Privacy Ordinance
• Artificial Intelligence fuels business transformations by learning about and knowing users and contents- Capgemeni
• How to treat the overlord clause in kuaishou user agreement?
• APT attack events
• Stuxnet APT targeting Iran’s nuclear facilities.
• Poison Cloud Vine (APT-C-01) : An 11-year attack on China’s defense establishment.
• Blue treasure mushroom (APT-C-12) : An 8-year attack on China’s nuclear industry and scientific institutions.
• CovenantSQL: CovenantSQL/CovenantSQL