"Cloud computing architecture technology and practice of" serial 22:2.4.7 media cloud

1952 people read comments(1) Collection Report

All rights reserved, without the written permission of HUAWEI, please do not reprint.

 

 

The overall architecture of the cloud computing cloud media solutions such as map based on2-35As shown.

 

picture2-35 Media cloud solution architecture subsystem

Media cloud solution scheme based on cloud computing platform of elastic computing cluster, the elastic storage cluster, distributed structured storage service, and distributed message queue service, broadcasting system of TV station provides efficient storage media, editing and cache to accelerate the calculation and storage resources hosting services.

Media cloud the cloud platform the key driving force comes from oriented public media of radio and television industry (including national and local television stations, radio stations) from the traditional cassette storage simulation to comprehensive digital media in the direction of development and evolution process, the demand of digital content and cloud data of large-scale intensive storage, processing and distribution. These requirements include: according to media gathering, editing, broadcast control, audio and video transcoding, media management. At present, domestic and foreign industry (such as Dayang, Sobey etc.) have launched a professional software for editing and broadcasting the media collection. Business processing in the application layer to solve the media content resources of these software focus. But the software on the bottom of the server and storage hardware based capability in ascension is still the lack of accumulation, such as: to enhance the efficiency in the use of server storage hardware, how to lower costs, provide greater storage space, processing capability andI/OThroughput bandwidth, reliability and disaster recovery capabilities, according to business demand allocation flexible resource pool etc.. Only the professional media management software and cloud platform closely, in order to achieve media data center infrastructure hardware resource utilization rate of increase, to achieve energy-saving emission reduction, to obtain the optimal hardware and software of the overall price.

Cloud computing platform is the core value for the media cloud data center is:

Mass storage capacity:Source from various sources audio and video media information content, persistent storage, because of the huge amount of data, storage cycle long (ThreeLast month), using the traditionalScale UPmodalRAIDControl storage in price is more and more difficult to meet the demand of large-scale cloud media.At the same time, a number of business media cloud, especially video/Audio media editing and online play, in between storage and computing resources of highI/OThroughput rateIOPS/MBPSPut forward moreHigh demand.Meet these needsDepends on the introduction of support stepless level extensionScale OutCloud storage. cloudstoragerealCost is proportional to capacity,I/OLaw of the Tuen MunAnd support across the cabinet, cross server, and across the data center disaster recovery reliability.

Cross business sharing of computing resources and storage resources sharing and balanced:According to media cloud application layer software of different business categories (production, editing, broadcasting, storage, management, etc.) at the same time resources occupy huge differenceFeaturesBy introducing the cloud platform, adoptUnifiedrealAutomatic telescopic resource pool platform supporting resources,The peak valley clipping and dynamic load balancing, bySOA Service/REST WebThe interface and application layer software to realize the automation of interactive resource management skills.,Average resource occupancy rateSurefrom20%Ascend to70%Above

Parallel computing and mass processing:According to the dynamic conversion between different formats of video media generally required (the cloudTS, H.264, MP4, AVI, RMVBDemand, etc.), due to the large size of the media file calculated capacity requirements of high density, by the maximum increase the encoding and decoding of the degree of parallelism, can make full use of available computing resources to divide and rule the media file, effectively shorten the encoding processing time required, to exchange resources for time and greatly improve business processing efficiency. Other such as advertising push to value-added services, but also depends on the parallel implementations of data analysis and processing platform based on consumer behavior analysis of historical data of the intelligence and automation of advertising strategy formulation and promulgation of.

Distributed caching and acceleration:forAlleviate the majority of Internet users hot on-demand multimedia content caused massive bandwidth demand and the limited wide area Internet connection bandwidth of contradictions.In the media cloud data center for media contentConductzoneandautomatic recognitionuserAccess hotspots,Hot content is automatically pushed to the distributed network nodescacheFor users to access the nearest centralized access to alleviate the bandwidth pressure. CachecontentCloud data centersourceContent meetingPeriodic synchronizationTo ensure that users see the contents of the latest content.

 

top
One
tread
Zero
Guess you're looking for
View comments
* the above user comments represent the personal views do not represent the views or position of the CSDN website
    personal data
    • Visit65081 times
    • Integral:One thousand two hundred and ninety-three
    • Grade
    • Rank:19217th name
    • Original65
    • Reproduced:3
    • Translation:0
    • Comments:17