DCN June 2016 - Page 19

Big Data & IoT Moore’s Law has played its part and increasingly dense processing power is reaching a point of diminishing returns. Data centre providers need to work smarter, not harder, to avoid being overwhelmed by IoT generated data. Artificial Intelligence (AI) and automated systems hold the key to managing the data overload. Removing the human element from data analysis and management allows for much greater speeds and power saving efficiencies. An effectively programmed AI system can reduce the amount of data that has to actually be processed and stored. It does so by dutifully monitoring all data streams, but only flagging significant variations from the norm (however this is defined) for further analysis. This allows the vast majority of data to remain ‘dormant’, reducing the processing power required to effectively manage a vast network of connected devices. This filtering step above the initial data collection system expands the capabilities of existing systems to monitor and analyse much larger networks than ever before. Disruption The Internet of Things is set to cause further disruption in the data centre industry, beyond the already substantial introduction of automated or software defined data centre resource management. A key factor of IoT is that it is populated by smallfactor devices running lightweight software, which can be installed in extremely remote destinations. As networks become increasingly detached, traditional, centralised data centre facilities will be unable to effectively manage and monitor them, resulting in a polarisation of data centre technology. Smaller ‘edge’ data centres will thrive in this new landscape of disconnected network clusters. Demand for processing power can be met as required with easily deployable, modular data centres installed locally. These jumping on/ off points for a larger network can provide localised computing power to monitor and process the data generated by the extended network. When it comes to edge data centres, the benefits of the modular approach are multiplied when the technology and capabilit W2&RFVWG6FRbFR6fVFFF6VG&RGVW2vW&RFǒFW6vV@FgV7F2֖GW&R&fFRFF6VG&R76RvFV6&vW"v6V7W&GBvV6V7FV@f6ƗG2r27VF&R6FF06&RWBFW&R2&V6FBGVR6( B&R6WBWB7FV@B6FRF7W'B6V7FV@WGv&GFW"r&VFRFW6RGVW2gW'FW"6RFFV"v6WGFw2vW&RB6( @f&RF7F6fVFVF7&Rf6ƗGVFW"g&FV6GvW""6V7FfGW'7V7FfRF27&V6rg&vVFFvB7VFRVBf"6fVFFF6VG&R76R&FW"G2&Pv6vRFFBbF&rখ7&V6vǒ6VvVFVBWGv&ࠤFF&f7FRfVVVBFB7FF&VFV0FgVFVFǒ&6RFRW7FpFF6VG&RWGv&2FRFV&FP&VBFF&f7BFRWvǐ&V7FFVB6fR&&"w&VVV@vfW&rFFG&6fW"&WGvVVFPURBFRU2VFW"G2WrwV6P2FR( &f76VN( FRw&VVV@26WBF&RFFVBVR'WB0ƖVǒF&R7V&V7BFVv6VvP'URVF6vW'2FR666FbFF6V6'27F&vR@&6W76r27&V6vǒb7&F6'F6RvFWBVVFp7V6f26G&2&WV&VBFVWBUP&VwVF2FF&6W2F6V@&6VB7F&vR67&VB7&70&VvW&6F7F26G&fVpFF&f7w2FW&R27&V6r&W77W&Rf 'W6W76W2FrW7FǒvW&RFFvVW&FVBg&7V6f2vVw&W20&VrG&6fW'&VBFǗ6VB"7F&VBF2&W77W&RgW'FW"V66W2FP'F6Rb6Ɨ6VBWGv&bVFvRFF6VG&W26V7W&VǐW6r&6W76rBG&6fW'&pBvVW&FVBFFvF7V6f0vVw&6&VF&W2ࠥFRFW&WBbFw226WBF6W6PgW'FW"F7'WFFRFF6VG&PGW7G'ࠥFW&^( 2FV'BFBBvFVƗfW"Wr( 66WGbFF( FvFǐ6V7Fr6V6'2BFWf6W2vG&6f&'W6W72B67VW ƖfR'f&֖rFFG&fVFV66rBWFFVB&6W76W0F&VGV6R67G2B&fFR&WGFW 6W'f6W2vWfW"bF2FFvVVBF&RFVB&WGFW"6'FW vFG&FFFF6VG&W267W'&VFǒ&fFR2&W7VBvRv6VR&6F67W"FRFF6VG&RGW7G'6WFrvW"v&RFWffVBg&6VG&Ɨ6VB6VG&W0F6VFvRf6ƗFW2BWFFv&RV'&6VB2FRV2`6G&ƖrBfFW&rFRWfW 7&V6rfrbBFF