Adhering to the “do the right thing, and not to do simple things” concept, policy data from the beginning of creation, god will use “privatisation deployment + SaaS” enterprise service mode, this is undoubtedly a difficult road, but through continued efforts in this direction, god ce data at present already have many scenes and deployment and environment, O&m capability of more than 1000 different clusters. At the same time, 71% of Shence customers have chosen private deployment solutions to consolidate and revitalize enterprise data assets while effectively ensuring data security.
As the first big data analysis and marketing technology service provider to support private deployment in China, Shence Data has built a strong competitive barrier in this field by virtue of its first-mover advantage and years of service experience.
Technical advantages: Standardized “one-click” deployment can be completed within hours
In the process of privatization deployment, different industries and customers can be described as thousands of faces; The enterprise operating environment is also varied in hardware configuration, hardware architecture, operating system, network conditions, etc. In order to accommodate these differences, they often accidentally fall into the pit of private deployment.
After experiencing a variety of complex and diverse deployment environments, Shence has developed a set of efficient and available private deployment system. 85% of customers can complete deployment within hours with the standard one-click deployment solution, which effectively solves the functional requirements of deployment, upgrade, service governance and monitoring of multiple application environments in all scenarios.
Hourly deployment, how does smart data work? The answer is standardization.
1. Unify the hardware standards and deployment process
First, the company developed a unified hardware standard to improve efficiency by unifying the physical environment. On the basis of full investigation and summary of past experience, the standardization design of cloud vendors, operating systems, disk performance, network performance and other key hardware environment is carried out.
Secondly, in order to improve the utilization rate of resources, god policy services are classified, and abstract the node, the node type three concepts, role, group, and attributes for each service, each service will belong to a character set, this will make the same kind of role deployed in the same type of machine, make full use of the machine resources.
In addition, Shenze has also formed a set of standard deployment service process, deployment service mainly includes physical environment standardization layer and cloud operating system layer, each layer hides its own management information from the upper layer. To be specific, standardize the physical environment before deployment and initialize the cloud operating system developed by Shence. The operating system provides a stable operating environment for service lines and facilitates the management of available resources.
2, operation registration system, modular design, so that deployment can be random
At the same time, private deployment is made easier and more efficient by means of operational registration and modular design that abstracts core concepts. For example, all service modules do not store any local data, and all data is stored in distributed storage. In this way, service modules can be quickly migrated when faults occur.
Considering the rapid development of services, deployment scenarios will also change, so when designing deployment architecture, Strategic pays special attention to conceptual abstraction and hierarchical decoupling, in order to quickly adapt to new deployment requirements. After the core concepts are abstracted, new deployment requirements are like building blocks that can be adapted to change.
3. A private deployment based on cloud native is coming soon
Through the improvement of the architecture and a complete automated tool chain, Shence ensured the reliability of the privatized deployment, but in order to completely shield the problems brought by environmental differences, Shence began to break the cloud native and container, after research and practice, Shence privatized deployment version based on cloud native is coming to the public.
The cloud native version is almost insensitive to the line of business, and the biggest benefit of the container is the isolation of the environment, which naturally avoids the environmental differences of a privatized deployment. At the same time, it will bring coupling between services and access problems inside and outside the container. Thanks to the great efforts made at the beginning of the design to prevent flying lines, the application of servitization, and the road to container transformation will come naturally.
When designing the cloud operating system, Shence adheres to the declarative design concept. As the platform provider, Shence allows the business line to only care about or declare the state it wants to achieve, rather than how to achieve this state. At the same time, in order to deal with the container challenge of stateful service, Shence developed a special stateful service management platform, which can not only flexibly manage and automatically expand capacity, but also connect with various services of cloud manufacturers, effectively take advantage of the flexibility provided by cloud manufacturers, and reduce costs for customers and facilitate operation and maintenance.
Most customers use computing resources periodically, which results in: Insufficient computing resources during peak computing hours, affecting user product experience; In the trough of computing, computing resources are idle, resulting in a great waste of resources. In order to solve this problem, the concept of tidal computing came into being. The so-called tidal computing is to supplement computing power through elastic resource pools during peak computing hours to improve user experience. When calculating the trough, recover the elastic resources and reduce the cost. At present, Shence’s elastic computing resources are based on kubernetes design, which well caters to the current cloud native trend.
Service advantage: Efficient and independent operation and maintenance system, can be as fast as SaaS iteration
Privatized deployment is not only a technology platform, but also an experience delivery platform. Shence data is experienced in providing complete private deployment solutions to customers from the very first day, and has delivered and implemented high-quality solutions to thousands of paying enterprises, and continues to run steadily and smoothly.
After serving thousands of customers, the private deployment of Shence Data has formed its own unique service advantages.
1, private deployment, SaaS flexible choice
In addition to the mode of private deployment, Some of Shence’s products also provide SaaS version, especially some partial real-time and online service products, SaaS version can get shence service team more timely, more comprehensive operation and maintenance guarantee. Privatized and SaaS environments seamlessly connect from cloud to cloud, allowing customers to choose flexibly based on their own needs.
Shence divides products into multiple applications according to their business functions, such as Shence user behavior data warehouse and Shence Intelligent operation. Each Application provides a series of interface-based services, and these interfaces are used by other applications.
As shence product applications interact with each other through interfaces, as long as the two Application interfaces are interconnected, they can be deployed in the same server cluster, separately deployed in two clusters, or even deployed separately in private clusters and SaaS.
For this reason, the magic strategy realizes the service discovery and access across the cloud. The business module only needs the service name to query the available port of the service. If the service is not provided by the Application in the cluster (for example, SaaS access across the public network), the connection will pass through the service path. On the service path, the system implements a series of functions such as authentication, data encryption, and audit.
2. Perfect operation and maintenance system
“Replacing human labor with machines” is a philosophy that Divine has always practiced, as well as in privatized deployment. We always hope to support 5 to 10 times multiple deployment solutions, automatic installation and upgrade of customer environments, and operation and maintenance without excessive increase in operation and maintenance manpower.
The privatized deployment of Shence Data includes a complete operation and maintenance system, deeply polished automated operation and maintenance tools, timely alarm, accurate positioning of anomalies, and help customers to operate efficiently and independently in their daily work.
For example, early in the design of a version of the product, you start thinking about the problems you might encounter when upgrading: how customers will accept the change, what the impact of the upgrade will be, how long it will take, whether it can be rolled back, etc. It is this way of working in advance and coping that allows private deployments to iterate as quickly as SaaS.
In addition, the divine policy monitoring system will automatically inspect, monitoring scope includes the machine level, internal and external modules; The emergency operation and maintenance event system wakes up automatically so that any exceptions can be seamlessly responded to.
In addition, in order to solve the migration problem of the privatized deployment, Shenpolicy has adjusted the migration solution several times, from the initial use of JDBC data export and import, to later reduce the available time after migration to 1 or 2 days. In the future, Shence will continue to optimize the data migration scheme, such as mirror migration, non-perceptive switching between new and old environments, and strive to complete the whole migration within hours.
Cost advantage: save hardware resources, help customers to achieve a balance in security and cost
Whether it is user data or business data, is the foundation of the enterprise. The privatized deployment scheme of Shence is combined with the fully open source SDK to ensure the complete transparency and security of data from collection to application.
On the premise of avoiding the defects of data integration security and stability between the public network and the internal network, the private deployment scheme has an open API interface, which can meet the flexible secondary development and provide strong guarantee for multi-system and multi-dimensional cross-analysis requirements. At present, many customers of Shence Data have realized precise push, literary works recommendation, CRM and other applications on this basis.
On the basis of data security, cost measurement is an important stage of enterprise selection of “private deployment” scheme.
In the face of increasing massive data and ensuring real-time query performance, The privatized deployment solution of Shence has low requirements on enterprise hardware configuration standards and quantity (CPU, memory, hard disk, and server quantity) and lower cost.
In the future, shence’s privatized deployment scheme based on cloud native will be unveiled. Through the self-developed stateful service management platform, it can connect with various services of cloud manufacturers, effectively make use of the flexibility provided by cloud manufacturers, and reduce costs for customers to facilitate operation and maintenance. God ce elastic calculation and elastic storage function can play in a variety of business logic with its strong technical capabilities, while save hardware cost, reasonable configuration optimization of enterprise computing resources, and provide related to “cold” data analysis ability, reduce the operation cost of enterprise, better help enterprises to achieve the goals of the business.
It is worth emphasizing that with strong scalability and scalability, the system can always maintain strong vitality in the whole application process. This means that even small and micro enterprises with only a single server hardware level can achieve low-cost, high-performance private deployment if they meet the requirements.
To sum up, on the basis of the reconstruction of Chinese data, the privatization deployment scheme is the only way for Shence Data to go, and it is also the inevitable choice for growing enterprises to carry out data construction. Shence Data interprets its solid barriers with its technical strength and service advantages.