1. All-around Streaming Architecture
The core of streaming system is to transmit media information in form of stream continually and steadily to the users end so as to ensure that users enjoy the service of online watching without having to wait and make full use of bandwidth resource of network.

However, the traditional streaming system merely transmits media in the form of stream instead of files while doing net transmitting. That leads to the disunity between the data sending model and reading, storing and dispatching model, as well as inefficiency coordination which in turn results in the reduced concurrent on-demands.

Therefore PowerInfo Streaming System creatively adopts the all-around streaming architecture: to use streaming system when sending, reading, storing and dispatching stream.

2. Distributed Architecture
The basic principle of CDN (Content Delivery Network) is to put edge servers at different areas of net and then redirect users' requests to the nearest service node from users in real time according to the integrated information including network flow, nodes linking and load status, the distance from users and the time needed to make response and so on.

But the construction of CDN system is based on conventional network files distributing and gradually increasing video distributing function. In the case of stream media with tens of thousands of nodes covering network, it contributes little to providing high quality service.

The distributed architecture introduced by PowerInfo Streaming System is a coordinated system oriented towards multi video nodes of media stream transferring. Every node possesses a balanced ability and cooperates with each other inside the architecture. The content of program flows dynamically in the whole system according to the using and operating rules. Distributed technology adapts better to the Broad Band Net environment in which information is dynamic and unsteady. Therefore it is an ideal technological mode to deploy large-scale streaming business.

3. Open Architecture
Most current streaming systems adopt closed architecture. For example, they only support certain media formats, transmission protocols and some specific operating systems and hardware platforms.

But as the fundamental business of broad band network, streaming system should have extensive applicability, support diversified media formats so as to adapt to different streaming businesses and different content suppliers, support different transmission protocols so as to adapt to different users' terminals, and support different operating systems and hardware platforms to adapt to different operators' infrastructures. Thus open architecture is essential to streaming system and shall be embodied on all levels and in all modules in the system.

1. Multithreading Non Semaphores Technology
PowerInfo Streaming System adopts original multithreading non semaphores technology innovated by PowerInfo. Using no semaphores between threads effectively reduces the system consumption resulted by synchronization and mutual exclusion of different threads, and in turn, reduces consumption of CPU. Latent system instability such as dead lock can also be avoided.

2. Multi Disk Store Technology
Conventional media streaming systems store video program in the form of files so that an individual file is completely stored in a hard disk. However, users' need is not average. The programs with high heat usually cause that the hard disks they are stored in become too busy. On the contrary, those with less heat leave idle the hard disks they are stored in, which is a huge waste of hard disk I/O resources.

Various forms of hard disk RAID functions exist today. But submitted to the capacity limit, neither can they cover multi disk arrays, nor can they ensure optimization aiming at video On-demand application.

PowerInfo Streaming System employs innovative multi-disk store technology (program being stored in stripped structure), dividing each video file into a number of data blocks which are distributed in different disks according to a certain calculating approach. When the system reads the data, it analyzes the information in the buffer and reads the program block through the pre-reading mechanism based on the current disk status such as position or status of magnetic head. This technology makes full use of all hard disk storage resources as well as I/O resources, not only improving the system's concurrence ability but also balancing disk load which helps to extend the disk's average service life

3. MAS(Multi Agent System) Technology
Distributed architecture is necessary to ensure large-scale program management, dispatcher control and load balance between video nodes, and MAS technology is the best choice for constructing it. The MAS consists of different agent with diverse function from various nodes, each agent in turn being a semi-autonomous system. Every function of stream media can be realized through coordination as well as alternation among agents. Constructing distributed architecture with MAS technology effectively ensures the avoidance of NP problem resulted from rapid growth of node number (NP problem refers to the incapability of arriving at computation results due to over complicated calculation amount).

PowerInfo Streaming System programs a variety of agents, including in-node load balance agent (as), media agent (ps), node management agent (ms), statistic agent (ts), dispatching agent (ss) and global load balance (gs) and so on. These agents cooperate with each other to realize various basic functions of stream media such as on-demand, storage and dispatch through algorithms such as step-by-step and agent collaboration. Furthermore, it can reach the highest efficient besides effectively avoiding single node failure.

4. Dynamic Stream Reconfiguration Technology
PowerInfo Streaming System adopts the original dynamic stream reconfiguration technology which ensures the realization of flexible denotation of various media formats, agile increase of a variety of media transmission protocols as well as the independence of media formats and transmission protocols, further offering effective supports for the opening architecture.

Through this technology, PowerInfo Streaming System is compatible with various media formats including MPEG1/2, MPEG4, H.264, WMV, RM, AVI, etc, and can rapidly extend itself so as to support new forms of media and transmission protocol.

5. 64 Bit Kernel Technology
As early as 2003, PowerInfo Streaming System was fully compatible with 64 bit operating system. It can carry out internal stored addressing as large as 2G which in turn ensures the realization of concurrence ability of 4,000 stream unit and intelligent content cache mechanism based on the computation approach LRU. As a result, its on-demand ability and I/O processing is far better than systems with 32 bit.

6. Pre Dissu Technology
In the process of transferring video information, the network particularly Wide Area Net is so intricate that the net bandwidth becomes unsteady, jittering or delaying, and transmission speed of disk I/O also jitters. All the factors mentioned above are likely to cause unsteady broadcasting of video data. To solve the problem and ensure best service of stream media, PowerInfo Streaming System adopts the Caching and Pre Dissu tech.

In accordance with current on-demands, PowerInfo Streaming System prereads some video data which then are placed in the buffer and output by streaming controlling threads. The intelligent Preliminary Reading computation helps the system improve the hit rate of the memory buffer content and reduce the I/O output of hard disk, thus achieve the concurrence performance of the system. On users' side, the player also installs a broadcasting buffer in which the data can ensure that the user watch smoothly during a certain period even if the net bandwidth happens to be abnormal.