Release Date: 8/25/2004 | Update Date: 8/25 / 2004thomas Rizzo
Microsoft Corporation
Summary: Tom Rizzo has launched his new column, which outlines the reason why the new "Longhorn" storage subsystem (codenamed "Winfs") was born. Winfs will help solve the role of data overload problems and his columns in the future. The content to be provided in a few months.
Welcome to read my new column "Winfs file"! This column is designed to help developers learn more about Microsoft® Windows® (especially code-based "Winfs" in the new version of the next version "Longhorn". In the subsequent sections of this column, we will understand the data models, functional sets, and programming examples of WinFS. The initial part of this column fully outlines the new features of WinFS and how to start learning and understanding this new technology.
Why launch winfs?
The technical industry is experiencing a more intense "perfect storm" - a storm that is combined by the trend and technology, which will drive the development and use of information to achieve the next breakthrough. This perfect storm is combined by three driving force: hardware improvements, digital data in quantity, and the surge in information management architecture and standards.
Hardware improvement
Looking back on the development process of hardware in the past few years, we can't help but feel surprised. Everyone knows that the Moore's Law - the number of transistors on the chip doubles every 18 months. But many people have ignored, the development speed of network bandwidth and storage technology even beyond the speed indicated by Moore. Both network bandwidth grow in an amazing speed interior and a single user of the global network through a home network. In the past few decades, the storage capacity on the server and client computer has increased significantly. In 1984, IBM PC launched 10 MB hard drives. Today, the 60-80 GB hard drive has become a standard configuration of a laptop computer. In the next few years, it is not difficult to imagine that the laptop will have 1 TB or even more storage. Along with storage at amazing speed growth, there has also been a problem in managing all data creation, storage, and searching for people every day.
Digital data
Most of the data used today is digital data. For example, when writing this article, I am not ordinary white paper, but the Microsoft® Word installed in a laptop computer. Now, email, electronic fax, digital media, calendar, Microsoft Office document, voicemail, and many other types of information are created and stored in electronically. A 2003 university of Berkeley Study (English) shows that the number of new information created only in 2002 exceeds 500 billion. 92% of this information stored in disk media (mostly client hard drives). The number of emails sent and stored in 2002 exceeded 400,000 trillion. Growing the power of hardware and software with computer connection, download, processing, and storage of more information than previously more information, and you will see the effectively managed digital data (the data we created by our own, and we are from others) How important is received by the received data.
Data standard and architectural data
The surge in data standards and architectural data is the last puzzle that is expected to consider new information management. In recent decades, many businesses and industry groups want to model data in their actual business processes in their actual business processes.
At the beginning, the computer cannot handle the complex process of modeling many business processes and automation. Now, with the improvement of computer operation and improvement of programming language and data technology, automatic implementation business process has become a way of simplifying operations, especially in corporate resource planning (ERP), customer relationship management (CRM), Salesperson Automation (SFA) and other types of enterprise applications appear. With the emergence of these new applications, the data suddenly became more complicated. But at the same time, the data has a better structure, and this is more useful for business. The database system is not used to store difficult-to-understand binary data or simple data models, but associate different types of complex data. For example, the ERP system understands the complexity of the total classification, human resources, and sales systems and modeling. Since the standardization architecture describes the data in the system, companies can ask for specific issues related to their services and can query this information in the system. However, cross-system work is still a problem in the industry. It is more difficult to integrate system within the enterprise, and integrated system between enterprises is difficult. This stimulates the development of enterprise application integration (EAI) suppliers because they can help customers span the gap between architectural data between heterogeneous systems. After recognizing this integration problem, the industry uses XML and XML Web Service as a way to help companies work between multiple systems and multiple organizations. These newer standards can help solve data integration issues and help share data within the organization and organizations. What is WinFS?
In order to welcome this perfect technical storm, Microsoft has put a lot of effort in building the next generation of Windows file systems (codenamed Winfs). Winfs product team follows the following three core principles in the process of revolutionizing the Windows file system: users can "find", "associated" and "operation". Let us understand the meaning of each principle, then discuss some technical details that make WinFS meet these goals.
Look up
According to the market and information technology investigation company IDC investigation, technical staff spend about 15% to 30% of time to find information. In the usual 8-hour working day (ie, a person works in the usual 8 hours of working day), you need to find information from 1 to 2.5 hours in any case. IDC also estimates at least 50% of Web search failed. Although the speed and accuracy of improving information look is one of the main goals of WinFS, it is still not enough to become an excellent search engine. Finding information has exceeded traversing and retrieving content. Instead, today's information has many meanings, such as the relationship between defined information (Winfs will provide this feature). Search may not be a default way for Winfs users to find most of their information.
In addition, within the past decade, the document system technology did not change much. Instead, some new data types (such as multimedia and new communication and collaboration), which are not just binary and simple metadata. The current file system does not know how to collect and find information in these new types of data. However, we are building WinFS to extend file systems, allowing them to include these information, which provide more information.
Association
Everyone understands the relationship between data. But today's software is not good in storage and use of these relationships. For example, I know that a particular document is introduced in a specific meeting. I know George wrote this document, and I know that another document written by George will also cause my interest. Now how do I find another document? I have to search by email and file system in the location I often accessed. With Winfs, Data Relationships will be built into the system and link the "all" data of the different "types" used to include custom application data you have written. Users can traverse these relationships and search data in a variety of ways. In addition, you can display these related items in a graphical form in an application that can identify WinFS so that users can better understand how their data is associated, whether the data is made by your application, others' applications or Built-in Windows program is stored. operating
So far, I have already introduced how to store all your data in WinFS and how to create relationships between these data in a more intuitive way. However, if this is all the things you can do with this system, then Winfs's innovation is only half. One key feature that people want to implement in any data system is to help them convert their data to useful information that can be operated. PC users spend a lot of time in one day in a digital staff acting as their data. You must manually sort, filter, classify, and stack grading. Winfs will provide a digital agent to help people transform from digital staff as digital data decision makers. One of the main ways of WinFS helps handle information overload is to eliminate redundant data activities that can be performed. Only information or important information should be displayed to the data user.
For example, you may want to know when an email that needs immediate reply from an author, this message has a relationship with a specific business document you are working. But you are not at your desk now, but is participating in a meeting and just with a mobile phone. You don't want to miss this e-mail and need to make sure you talk to this person as soon as possible. Introduction to the other party's conversation requires a 30-minute telephone to exclude all issues in business documents (the deadline for this phone is before the end of today).
Winfs Rules is the built-in component of the system that allows you to inform the system how to handle, order, and deliver data. By using Winfs Rules, you can create a rule of data and data relationships. In addition, Winfs Rules can also work with other Windows applications to give you a notification (for example, sending paging to your phone). Finally, Winfs Rules can also help arrange a conference call by viewing the idle time in the day by viewing the calendar information. Here, the rule can automatically create a meeting in the next idle time. Integrated Winfs Rules technology allows data to be stored in WinFS (even copied to WinFS) to make it active data. This activity data will be notified to you when it enters the system, thereby helping to make better business decisions. This technique can help ease the information overload by automatically handling many of the tasks that we now manually perform data. technical background
Now, we have discussed the basic principles of WinFS, and we will introduce some of its core technologies. In this column, I will provide high-level overviews for each technology. In the next few months, this column will introduce each technology and show the developer how to use the new WinFS technology to establish an application.
From a technical point of view, Winfs consists of 5 parts: core Winfs, data models, architecture, services, and APIs. Figure 1 shows a more detailed view of the structural blocks of these parts.
Figure 1. Core Winfs Structure Block
Core Winfs
The core WinFS consists of the core service of the file system. The core WinFS is deemed as the foundation, which includes operation and file system services. Its example includes security, easy management, Win32 file access support, import / export, quota, etc. Data model
The data model is not included in the core service, which provides some technological innovations mentioned earlier, including basic terms, relationships, and extensions and relationships.
Architecture
If there is no built-in architecture, Winfs will not be better than existing file systems because WinFS does not understand data in a richer way to understand data or provide more structured ways to process data. The WinFS architecture includes architects of daily information (including documents, email, tasks, media, audio video, etc.), as well as system architectures including configuration, program, and other data related to system.
service
Synchronization and rules belong to WinFS's service category. These technologies are located in "top" of WinFS, which can provide you with extension features other than system infrastructure. The synchronization feature allows you to synchronize the Winfs system over the network and you can establish a synchronous adapter to synchronize Winfs to other systems. For example, you may want to synchronize contact information from your CRM system to WinFS so that you can use this data to be associated with other data in WinFS, or use this data offline from WinFS. The sync adapter can be two-way, so all changes to the data in WinFS can be synchronized back to other cooperative systems.
API
As a developer, you will write an API. Winfs includes a feature-rich API, which is part of the entire WinFXTM programming model in Longhorn. With the WinFS API, you can write different structural blocks of the WinFS system, including data operations, rules, synchronization, and data models. in conclusion
In the next few months, there is a lot of content to be introduced because of the beginning of deeply exploring the different technologies of WinFS. At the beginning, you will want to make a clear understanding of other longhorn cornerstones, especially "Avalon", because many applications that will build will be based on Avalon. It is recommended that you see Chris Sells overview of Longhorn cornerstone in its First Longhorn Foghorn Column (English). In addition, let us prepare for WinFS: it is the future of information-driven applications.
Winfs file
Thomas Rizzo is the supervisor of the Microsoft SQL Server group. TOM uses spare time to write programming in Microsoft Press, and help customers in the Microsoft News Group and occasionally update their blog (he should update more often!). You can contact Tom by email, his email address is: Thomriz@microsoft.com.
© 2004 Microsoft Corporation All rights reserved. all rights reserved. Use the specification. Go to the original English page
