COMPARED TO DISK SYSTEMS , TAPE CONSUMES MUCH LESS ELECTRICITY AND GENERATES ONLY A FRACTION OF THE HEAT .
EDITOR ’ S QUESTION
INES WOLF , PRE-SALES MANAGER FOR CENTRAL EUROPE , QUANTUM
ith continuously
W increasing data volumes and simultaneously rising costs for electricity , companies are faced with an enormous problem : how can data centres achieve their net zero goals during an energy crisis ?
Historically , green and cheap have often been mutually exclusive . In the search for a solution to the problem , one inevitably comes across the biggest energy consumer in the data centre : high-capacity storage , which traditionally relies on energy-hungry HDDs .
To reduce the energy consumption of data storage , and consequently the CO 2 consumption , there are in principle several levers that companies can use . Considering that 90 % of all generated data is rarely used again , there is a huge potential for savings . Tape storage plays a key role in achieving those savings . Compared to disk systems , tape consumes much less electricity and generates only a fraction of the heat . It consumes virtually no energy when not in use . On top , those energy savings increase exponentially with the amount of data stored . For example : at 0.5 PB of data stored , one can save 50 % of electricity with tape . At 50 PB , savings of over 95 % are possible as efficiency of scale applies .
However , moving all data to tape to save kWh and CO 2 is not feasible . Data centres must always guarantee the highest performance and availability of the data they hold – something that can hardly be achieved with tape only . Looking at a good balance between high and low-performance data storage , a twolayered approach would fit : a super-fast NVMe-flash-tier for high-performance workloads and a much cheaper tape tier for less frequently used data . Both use much less energy than HDDs , but the question is not only about the right balance of these two layers , it ’ s about how to meet an organisation ’ s dedicated requirements and how different storage layers can be efficiently managed on a software layer .
HDD storage still has its justification . Organisations must manage huge amounts of data only by one single integrated data platform instead of a patchwork of various solutions , and because it is practically impossible to manage huge amounts of data manually , this data platform needs to be AI-driven . AI allows the optimisation of resources as fewer human man-hours are needed to complete repetitive tasks . AI systems can automate complex tasks that would be time-consuming or even impossible for humans .
When discussing the costs of a hybrid storage solution concept , it is important to also consider the sustainability of the individual storage components . The typical life cycle of a disk storage
COMPARED TO DISK SYSTEMS , TAPE CONSUMES MUCH LESS ELECTRICITY AND GENERATES ONLY A FRACTION OF THE HEAT .
system is about five years , while a tape storage system is usually used for more than 10 years . With this duration of use , not only the sustainability in terms of operation comes into play , but also the sustainability in terms of the manufacture of the storage systems .
Net zero in the data centre is possible . With the right strategy and solutions , organisations can save huge amounts of energy and CO 2 by optimising the data centre ’ s biggest energy consumer : storage . The savings can be used to offset the higher cost of flash storage and the more expensive green power . This approach is also future-proof as the savings scale exponentially , which will help manage the ever-increasing amount of data created and stored . Compared to traditional HDD storage , organisations can save up to 95 % energy for data storage if they move as much data as possible to tape . www . intelligentdatacentres . com
33