New Architecture for Enterprise

After gaining a basic understanding of the functions and features of Eight, let us now explore its potential application scenarios. Specifically, what can Eight be used for? What problems can it solve?

Common Issues in Enterprise Applications

Enterprise applications differ significantly from internet applications, with some prominent pain points such as:

• Enterprise applications often have complex and unique business logic, requiring substantial customization, and the business itself is continually evolving, necessitating frequent customization and maintenance.
• Enterprise applications are typically distributed and dispersed throughout the entire enterprise system, existing in various departments, branches, and even discrete or edge nodes. This contrasts with typical internet applications, which are centrally deployed in controlled core facilities (such as public clouds).
• The standardization level of enterprise applications is relatively low, and the degree of reuse is poor. Unlike internet applications, which can spread development and maintenance costs across a large number of users, customized enterprise applications are often used exclusively by the enterprise itself, which has limited capacity to bear high R&D costs.
• Most enterprises do not have the advanced information technology conditions (such as virtualization platform and private clouds) and lack sufficient technical talent. Advanced internet technologies are difficult to implement and even harder to manage and maintain.
• The computing infrastructure environment within enterprises is complex, with various edge nodes and information silos, running numerous legacy systems and outdated equipment, with unstable network environments and stringent security policies. Enterprise application systems need to be deployed and extended in such environments.

Specifically, in China, due to national conditions, policies, and long-standing technical traditions:

• Chinese enterprises prefer privatized, localized solutions, especially large enterprises and state-owned enterprises, due to policy and regulatory requirements. The security and privatization of business and data are core demands. This means that all the above issues cannot be alleviated through SaaS-based internet services.
• The development across regions in China is uneven, with technical personnel concentrated in a few cities, while enterprises, especially large state-owned enterprises, have business spread across the country. Generally, enterprises find it difficult to build technical teams, let alone establish such teams in their branches.

National Enterprises with Extensive Branch Networks

Such enterprises are quite typical, such as banks, securities, financial insurance, petroleum and electricity. The characteristics of these enterprise systems are:

• They have private clouds and excellent infrastructure, with the main systems centrally built on private clouds.
• They have a quality R&D team in central cities, capable of developing and maintaining centrally built systems.
• Numerous branches across the country mainly use SaaS systems on private clouds.
• There are still many applications that cannot rely on the network and SaaS (such as various offline business systems), which need to be managed, upgraded, and maintained, requiring significant investment to build a national technical support team.

Taking banks as an example, their information system resource structure is roughly as follows:

This model is suitable for the advanced internet-oriented technologies that have developed in recent years, such as big data, virtualization, private clouds, etc. These enterprises also have sufficient capacity to build a set of hardware and software facilities and talent teams.

However, the main pain point for these enterprises is the high system maintenance cost for branches, especially when maintenance involves software and business. A small number of business R&D personnel are concentrated in core cities, while problems may arise in various locations. Local maintenance personnel do not have sufficient capability and authority to solve problems. Therefore, the current approach is to migrate business to the cloud for centralized management as much as possible, but the existence of offline business is an unavoidable reality.

If Eight’s solution is adopted for localized applications in branches, such enterprises can build a centralized application distribution and maintenance system. Business systems can be centrally developed and deployed at central nodes, with branch personnel connecting to the central node’s services in a simple manner. The central node can deploy and upgrade systems locally. When maintenance is required, branch nodes connect to the central system, allowing R&D personnel to track, debug, and maintain these remote local systems online. This is very helpful for banks to promptly resolve node failures and reduce maintenance personnel costs.

Compared to previous solutions, the advantages of Eight infrastructure are:

• Reduced maintenance costs, facilitating the construction of a reasonable system-personnel structure, enabling enterprises to use a small number of high-quality technical resources to solve remote problems across various locations. The objective reality in China is that high-quality technical talent is concentrated in core cities, and deploying high-tech personnel in other regions is not only costly but also lacks available talent. Eight elevates the central distribution and remote control of systems to a higher level, making it possible to maintain and manage with a small number of high-quality personnel. This not only reduces personnel consumption and alleviates the talent shortage in small and medium-sized cities but, more importantly, allows each enterprise branch to enjoy high-quality technical services.
• Improved problem and fault response efficiency, allowing faults to be eliminated at the first occurrence, which is particularly significant for enterprises requiring timely business responses, such as banks, finance, and telecommunications.
• Easier access to the site, analyzing causes, identifying faults, and improving product quality. Eight can remotely enter a runtime system, observing the capabilities of each functional component, greatly facilitating the ability to troubleshoot real-time or random faults.
• Making many businesses no longer reliant on SaaS possible. The current widespread SaaS adoption is not entirely due to business requirements but is largely constrained by technical means. Past technical means required centralizing business on private clouds to facilitate centralized management and upgrades. However, SaaS applications have inherent drawbacks, such as over-reliance on the network, inability to withstand network node fluctuations, over-reliance on the center, making the central system a bottleneck, and excessive centralization and uniformity, making it difficult to adapt to different branches’ local conditions. Additionally, many businesses may inherently be local applications (such as many offline-online combined applications). Eight provides a new option for such situations. Applications distributed by Eight are essentially local applications that can run without relying on the SaaS network environment while ensuring maintenance, upgrades, and management of these applications. Enterprises can flexibly choose solutions based on business characteristics rather than technical conditions in the future, providing more robust and locally adaptable applications and services.
• Eight is relatively easy to implement in such enterprises. These enterprises already have their privatized network systems and cloud platforms. Eight’s business release can be built on the central cloud platform, and branches only need to deploy the Eight seat according to business needs and connect to the cloud platform through a trusted network, allowing the enterprise’s software R&D to issue business.

Managing Edge Systems

Another common scenario is in industrial manufacturing. With the development of the industrial internet, numerous traditional or emerging enterprises are beginning to have various systems, often widely distributed, with complex environments, and the stability and reliability of the systems are challenging to guarantee.

A typical application scenario is smart cars. For car manufacturers, managing and reaching millions of cars worldwide is a challenge, especially since these cars may span decades of production, with a myriad of systems and applications deployed on them. This issue will become more severe in the upcoming smart car era, where cars may become the next daily life gateway after smartphones. However, unlike phones, people do not frequently replace their cars or upgrade their car systems. Modern car systems mostly support OTA (Over-The-Air updates), but the scope and cycle of OTA are hard to estimate. After all, OTA requires customer cooperation, takes a long time, and cannot be performed while the car is running. Cars often disconnect from power and the network when stopped, and updates are usually only possible during maintenance at 4S shops.

Therefore, this is not an always online solution.

Consider a scenario where user behavior on car systems needs to be collected. The content and format of the collected data may need to change continually based on manufacturer requirements. To standardize data formats and collection content, manufacturers would need to OTA millions of cars worldwide simultaneously, which is nearly impossible. Otherwise, manufacturers will have to endure compatibility with various data formats and content, and new collection businesses will never be implemented, severely impacting the manufacturer’s application needs. As a data-driven growth strategy, such collection needs may change frequently, with each change presenting the same challenge. This situation can be addressed with the Eight solution, as follows:

Compared to previous solutions, the advantages of adopting Eight are:

• Eight itself provides always online application deployment capabilities. Without affecting basic functions (i.e., without requiring OTA), manufacturers can have a lightweight update container that can be arbitrarily controlled and managed, loading high-control and management business into it.
• As mentioned in the previous section, unlike OTA updates that are hundreds of MB or even GB in size, applications under Eight are very small, often only a few hundred KB for the entire system; updating a single component is often only a few to dozens of KB. Thus, even in poor network environments, the entire push can be completed in milliseconds to seconds under 3G or 2G networks (essentially a simple HTTP request). These operations can be performed as long as there is a network, providing very practical update capabilities for cars, which are nodes with unstable environments.
• Eight deploys a local application, meaning that apart from deployment and maintenance requiring online connectivity, the application itself does not necessarily need a network. This is particularly useful for edge nodes with unstable network environments (such as car systems). In this case, the designed application can be more flexible, without considering continuous online requirements (for example, for data collection, continuous local storage can be used, with batch updates when connected to the network).
• Eight is designed with compatibility in mind, supporting Java 1.6 at a minimum and running on all devices with Android 2.3 and above, covering most of the currently running smart terminals.

Managing Discrete Nodes

Another scenario is in industrial enterprise manufacturing. There are numerous discrete systems and nodes within enterprises. How can these islands be connected into an integrated system for centralized monitoring, management, upgrading, and maintenance? How can they be integrated with the enterprise’s future intelligent production and manufacturing systems?

Currently, there is no solution to this problem. Whether it’s private clouds or OTA, neither has considered providing solutions for various devices, terminals, local application systems, and scattered businesses under various environments within factories. These areas have been forgotten, unable to deploy k8s or use Android.

However, Eight provides a solution:

After adopting Eight:

• Eight is designed with traditional system compatibility in mind, capable of running in any legacy device environment (Windows compatibility extends to the 1990s’ WinNT4 and Win9X, Linux compatibility extends to the 1990s’ kernel 2.4.x. For Eight’s design and compatibility capabilities, refer to the documentation).
• Eight is easy to deploy, occupying minimal resources, suitable for continuous operation in various resource-constrained systems.
• Traditional enterprise information technology was concentrated in the early 2000s, with many applications based on Java. Using Eight can naturally integrate with these systems.
• Java’s rich libraries and extensions facilitate the control of discrete nodes and applications. For example, for applications developed in C, JNI can be used to integrate, monitor, collect data, and manipulate them.
• Unlike remote login management solutions like SSH, Eight does not need to penetrate multiple subnets through security policies to reach device nodes, especially when these discrete nodes are deep within the enterprise. It also does not require nodes to be continuously online, which is particularly useful for devices under various security management. Nodes can connect to the central control node through multiple firewalls and limited networks, becoming part of the controlled system.
• Once all discrete nodes are connected as an integrated system, enterprises can extend their control tentacles to every corner, monitoring, maintaining, upgrading, and deploying applications to systems across the enterprise, collecting and integrating various data, and connecting business processes, forming part of a modern intelligent enterprise.