Is the data stored in the cloud really safe? The future of cloud services will be like this

Time:2019-9-29

In this era of information explosion, data storage is closely related to each of us. From punch cards to floppy disks and hard disks to centralized cloud storage services, human beings have never stopped on the road to more convenient and effective data storage. Will the most popular centralized storage service be cheaper, faster and more private in the future? In the follow-up articles, we will start with the problems in today’s market and take you slowly to appreciate the charm of centralized storage.

How big is the storage market?

When we talk about a particular market, size and trends must be the two most important points. From the figure above, the total volume of hard disk shipments is estimated to reach 1120 EB by 2020, and the trend has been rising. So, what kind of problems exist in such a huge market?

What is the biggest problem in the current market?

Here, we first introduce the storage industry. Storage is divided into enterprise storage and personal storage. The best examples of enterprise storage are AWS’s S3 and Aliyun’s OSS. Personal storage is a familiar network disk for backup and synchronization functions, such as Microsoft’s Onedrive, Apple’s iCloud, Google Drive, and third-party Dropbox.

Whether it is enterprise service or personal service, the most popular or centralized scheme. Businessmen build their own data centers, buy bandwidth, purchase machines, deploy software, and build a complete service product, which is called centralized storage.

Although centralized storage is popular, it is not perfect. It has the following problems.

  1. Centralized storage, with a clear center, has high security costs and is more vulnerable to attack. As we all know, for hackers, as long as they break through this clear center, they can get all the data, even tamper with it.
  2. For operators with insufficient strength, there is a risk of termination of services at any time. Previously, China had a period of popular online warfare, and this is not a price war directly initiated by start-up companies. The participants had 360 nets, Jinshan nets, Sina Weibo, Baidu nets and so on. Over the past few years, most of the disk business has been shut down, and a large number of users have lost their storage data.
  3. Personal data is no longer owned by individuals, on the contrary, the platform has all the data. For their own benefit, some platforms will use these data as training data samples for AI in-depth learning, and ultimately seek benefits in other fields. These data have certain commercial value after disguised analysis. Businessmen can use these data for user portrait analysis, so as to accurately guide users’behavior habits and mental deduction, and ultimately induce users to buy goods, thus generating value. Therefore, as long as the data is centralized, the user’s data rights will not be guaranteed ultimately.
  4. There is a risk that personal privacy data will be censored. First of all, we believe that the review of illegal data is conducive to social and economic stability. But for the censors, when censoring, because they are not sure what data is illegal, what data is not a problem, only to review all the data once, which leads to whether there are problems in the data will be censored, and will inevitably affect the legitimate privacy of legitimate users to be seen, or even. The possibility of being leaked.
  5. Bandwidth transmission costs are very high. Note that what I’m talking about here is bandwidth, not storage. Whether uploading or downloading data consumes bandwidth, so bandwidth is not negligible and is more expensive than storage cost. And centralized storage services, in order to ensure that users everywhere have a good experience, they will deploy the data center to the backbone network, which is the most expensive in any country.

Because centralized storage has such a problem, we consider using decentralized storage scheme to solve it.

Why Decentralized Storage?

Decentralized storage is no longer about building data centers, purchasing bandwidth, purchasing machines and hard drives. Instead, it is about building a fair, effective and transparent set of rules to mobilize the storage capacity of the world and the resources that are effectively connected to the Internet. Then, through a unified distributed software system, all resources are effectively linked up to provide a stable service. This is in line with the essence of P2P technology, that is, to build stable services through probability rule on relatively unstable infrastructure.

What are the advantages of decentralized storage?

  1. Low transmission costs:Centralized storage can only deploy services to backbone nodes with high bandwidth cost in order to preserve user experience in many places, while decentralized storage can make full use of cheap MAN bandwidth, even idle home bandwidth of users, which can greatly reduce bandwidth cost and more specific content, see my article. The previous article, “Why PPIO Prices Can Be Lower than Aliyun and AWS”.
  2. Absolute security:The security mechanism of decentralized network is different from that of centralized network. The security of centralized network depends on the credit of centralized operator. Its core storage mechanism is closed and open source. The centralized operator organization has the same authority as users to view, modify and delete data. The security mechanism of de-centralized storage is different. It is decided by mathematics, cryptography and network mechanism. The code of its core storage mechanism eventually appears in the form of open source transparency. Data viewing, modification and deletion are determined by the user’s private key. As long as the private key is not stolen, the data is secure. Like Bitcoin, as long as the private key is not stolen, all funds are secure.
  3. Never stop:For the users of de-centralized storage, there is no need to worry about the possibility of operator shutdown, because the final de-centralized storage belongs to the users, the community, and not the company. Specifically, as long as the network can run, it will eventually emerge as Bitcoin in the form of fully open source, because ultimately there is no center, as long as there are people willing to deploy nodes, the entire network will automatically connect all nodes, thus providing valuable service network.

To sum up, decentralized storage is a potential platform with commercial value. Note that the value mentioned here does not mean that the higher the price of the currency is, the more valuable it will be; nor does it mean the number of beliefs and consensus; it refers to the value generated by real needs and satisfying needs in the process of real commercial landing.

It is for this reason that we launched the PPIO project, which is a de-centralized storage and distribution platform for developers to make data cheaper, faster and more private. The official website is https://pp.io.

Here, I say PPIO is a platform, so what is a platform? The so-called platform has at least two roles, one is supply, the other is demand, which facilitates the matching between the two roles and eventually forms a transaction, which is what the platform does. Moreover, the value of the platform is determined by the orderly and rapid growth of all parties on the platform.

What are the important participants in PPIO Decentralized Storage Platform?

There are three kinds of roles in PPIO de-centralized storage platform:

  1. The first category is users, who use storage services as the demand side of the platform.
  2. The second category is miners, which are storage nodes. As the supply side of the platform, they provide storage services and bandwidth services generated by uploading and downloading.
  3. The third category is service nodes, because the entire platform also needs some nodes to provide special, organizational services, such as index retrieval, supervision certification, block, and the implementation of the Oracle machine.

So how should PPIO network start in the early stage? Most block chain projects are started by pulling the supply side. Next I will focus on the analysis of PPIO cold start thinking. Are these three types of roles pulled at the same time? Or in the early days, pull the demand side or the supply side first?

How does the PPIO cold start process work?

Since it is a platform, there are early cold start strategy issues. So is such a market suitable for unilateral, bilateral or multilateral start-up in the early stage? This is an interesting question, we might as well make an analysis.

First of all, when the service node starts cold in the early stage, it does not bear much load, so it can deploy the server itself to assume the role of the service node. After users and storage nodes are loaded and software programs are verified stable, they can be gradually opened to third parties to deploy service nodes together. This is also the first strong center, then weak center, and finally the process of going to the center. See article PPIO’s three stages: “strong center”, “weak center” and “go center”. Therefore, service nodes do not need to focus on early operation, try to pull.

Then we discuss the user node and storage node, which is a classic topic. Its essence is the topic between supply and demand. First we do some analysis of other Internet platforms, tremble, micro-blog, fast hand are pulled by the supply side. They are all competing for the creative side of content. The creative side is personalized, non-standardized, and good content can bring great value. This is in line with their business logic, but these are supply-side and one-sided. Pulling. The cold start of Jingdong Mall is not driven from the supply side, but from the demand side, because Jingdong sells standardized products such as digital 3C. In the early stage, it used independent operation, focusing on the pressure of growth on users, which is driven by the demand side unilaterally.

For our PPIO Decentralized Storage Platform, the essence of the products it sells is storage and bandwidth. Apart from geographical factors, all of them are standardized products, without any creativity or special parts. So we need to adopt a strategy similar to that of Jingdong, that is, to adopt the demand-side pull mode in the early stage, that is, to focus on the development of user use.

The reason is not difficult to understand, do a simple deduction. If we use supply-side pull, but the demand is insufficient, soon miners will find that they can not obtain stable and sustainable income on PPIO platform, then miners will leave. On the other hand, if the demand side pull is adopted, the users will grow in the early period when they focus on operating users; after accumulating a large number of users, the consumption of the entire PPIO platform will increase. Miners provide more and more stable services on PPIO platform, so as to obtain more and more stable benefits. Only in this way can miners provide services in PPIO without stopping or leaving. In the long run, the supply side will become more and more stable, and the demand side will be stimulated to use more and more. Finally, the process of cold start-up is accelerated. Please refer to the article “should ppio first pull supply or demand?”

Since this is a platform, in addition to pulling the demand side to complete the cold start in the early stage, we need to establish an effective market mechanism, that is, storage leasing market and bandwidth leasing market. Matching and matching in quotations between users and miners. Users hope that the unit price of service is getting lower and lower, while miners hope that the unit price of service is getting higher and higher. Finally, they hope to achieve a balance between buyers (users) and sellers (miners) in different regions.

Now imagine PPIO platform has gone through the process of cold start and has a certain scale, then PPIO will appear other Internet platforms often encountered problems? I will analyze this part in detail later.

What difficulties may PPIO platform encounter?

First of all, I would like to talk about what problems other Internet platforms may encounter when they become bigger. Taking Taobao as an example, it has had the following problems:

  1. The problem of competition on the same side:There is fierce competition between businesses of the same category. It is difficult for new entrants to open stores in Taobao.
  2. Information overload:When the number of goods is very large, it is very difficult for users to choose goods.
  3. Decline in service quality:As more and more commodities have gone through fierce price competition, the quality of commodities becomes more and more difficult to guarantee. Finally, the quality of service can only be maintained by cracking down on counterfeits.

Now let me talk about how PPIO considers these issues.

1. Homologous Competition

The main consideration is that there will be fierce competition among storage nodes. The result is that it is impossible to run full storage, because there are lower-cost, more stable storage nodes in the PPIO network to provide services. In the same region, the same service commitment (mortgage), the same physical structure of the hard disk material, our algorithm will give priority to the use of lower prices, better stability of storage nodes. I think this algorithm design is correct because we provide standardized services on the supply side. It is precisely because the algorithm will give priority to guaranteeing lower prices, more stable storage nodes are easier to receive orders, thus establishing an effective competition mode on the supply side, and promoting storage nodes to achieve more stability and lower prices. Imagine that it’s not like today’s Bitcoin. If you’re not a professional miner, you can get very cheap electricity, or you’ll lose money digging Bitcoin. The root of all this is the relatively fair economic mechanism. If subsequent orders are to be obtained on the PPIO network, try to get cheaper networks, storage and electricity.

2. Information overload

Information overload usually occurs on the user side. The reason is that there are more and more storage nodes on the supply side, and the PPIO platform will show more and more storage nodes to users, resulting in users do not know what kind of storage nodes they should choose. So we give the selection mechanism to the algorithm to solve. Different algorithms can make different choice logic, but the essence of de-centralized storage is to build a standardized service. Technically, we recommend the default stock transaction matching mechanism similar to stock exchange. That is, in the case of the same geographical location, service commitment and storage physical structure, price is the first choice factor, and quality of service is the second choice factor. In this way, we can quickly make a choice, find storage nodes, and automatically sign storage contracts (i.e., transactions). In short, PPIO solves the problem of information overload through technical means.

3. Decline of service quality

Faced with the supply-side competition mentioned above, the storage transaction price of PPIO platform is getting lower and lower. Will storage nodes consider reducing the storage quality and price, such as using low-quality and low-cost hard disks? This is why PPIO designed Token’s economic mechanism. All transaction and service execution details of PPIO will eventually be written to the block chain after the certification. PPIO designs an economic penalty mechanism for storage nodes that are not fully proven on time. For storage nodes, if they use bad hard drives and bad networks, they will have problems when they finish their services. Storage nodes will be penalized for problems, and mortgages may be penalized. For storage nodes, it is not worth the loss, so they will find their own way to ensure the quality of service. Simply put, PPIO is an economic mechanism to solve the problem of poor quality of service of storage nodes.

See here, do you have a clearer understanding of the feasibility of decentralized storage as a valuable business platform? If centralized storage is the mainstream of current storage services, then de-centralized storage must be one of the most valuable commercial platforms in the future. Are you ready?

For more information about PPIO, go to the official website: pp.io

Recommended Today

Detailed explanation of spring cloud ribbon load balancing Shenkeng that cannot be connected through the server name

First, the problem. Take the Eureka cluster and the client calls the service through the ribbon. The ribbon side reports the following exceptions java.net.UnknownHostException: SERVICE-HI java.lang.IllegalStateException: No instances available for SERVICE-HI java.lang.IllegalStateException: Request URI does not contain a valid hostname: http://SERVICE-HI com.netfix.discovery.shared.taransport.TransportException: Cannot execute request on any known server The spring cloud version is messy, […]