Chinese telecomms company Huawei has created a news Data Center 3.0 (DC 3.0) architecture, based on resource pooling, software-defined and all-optical interconnecting technologies. The new architecture developed in Huawei ‘s Shannon Lab is intended to improve the real-time data processing ability of data centres in the future in anticipation of the increased traffic from cloud applications, big data and the Internet of things (IoT).
Huawei claims that by using flexible real-time software and hardware resource allocation, there is a “significant improvement” in the ability of data centres to process data. In addition, the DC 3.0 architecture has a flat configuration and the capability to expand flexibly, which Huawei believes will reduce data centre costs.
Over the next 10 years, Huawei expects that average personal data consumption will reach PetaByte levels, 1,000 times larger than Terabytes, the current data consumption level. The increase of data consumption will result in conventional data centres facing challenges such as real-time data processing, on demand resource allocation and massive systems energy efficiency.
The key innovations in DC 3.0 are pooling computing and storage resource via disaggregation. Parallel computing can improve the processing ability of large scale data centres. The silicon photonic interconnections with fine granularity resource management enable the architecture to meet the high bandwidth and low latency requirements of data transmission. Also, processors will be able to achieve the tasks assigned by the heterogeneous processor architecture.
DC 3.0 also looks at the problem from a data-centric architecture rather than the conventional computing-centric mode. In this architecture the data can be allocated to specific processing units, making it flexible to manage.