Recently I read an article on cutting-edge research demonstrating the feasibility of data storage on media for a million years! A few years ago this might not have resonated with me as a thermal engineer working to cool electronics, but now it does. Let me explain why.
Research & Development on IT equipment thermal management over the past few decades has justifiably focused on microprocessor (CPU) based systems. This may be explained by increased CPU transistor density, as per Moore’s law, hence rising heat fluxes. However, in parallel, the growth rate of storage hardware that complements the compute engine (i.e., microprocessors) has been dramatic in recent years, not only in data centers that support internet based products and services (e.g. social media, photo/video hosting, utility/cloud computing, e-commerce), but also in the governmental, medical, and industrial sectors. While microprocessors come and go with new electronics hardware, the expectation for how long the data need to be stored for can easily range from decades to a lifetime to potentially centuries (e.g., in the case of historic and scientific data). As humans with intricately embedded electronics, we generate and use data at an ever increasing rate. This should not come as a surprise to any of us, engineers or otherwise, based on our own data storage needs at work or at home. On a personal note, I own several storage devices ranging from CDs, consumer disk drive devices, and DVDs. The number of storage devices has continuously grown rapidly over the last few years, especially with the addition of so many photographs of my daughter since her birth.
So, why should this matter for a thermal engineer in the electronics cooling field? In addition to cooling traditional electronic devices and system, some of us are responsible for reliably cool storage devices and systems (and data centers). Storage devices could be made of magnetic media (tapes or rotationally spinning disk drives) or silicon (flash) or something else. As an example, a common form factor for hard disk drive with spinning magnetic media is 4” by 6” by 1”, with a power in the range of 10 W and maximum case temperature of approximately 60oC. By itself, this thermal management problem is a piece of cake for a thermal engineer, but consider hundreds or thousands or millions of such drives closely packed with each other, with fan-based air cooling doing the job. Now consider that some of the devices store “hot” or active data, and others store “cold” or inactive data. Each of these storage tiers have different cooling needs over their lifetime. I anticipate that this area of thermal engineering will increase in importance in the coming years, and I thus wanted to use my editorial capital for this issue to discuss this topic with the ElectronicsCooling audience.
For this edition of the magazine, I am excited to offer the readers several interesting topics such as justification for embedded liquid cooling for very high performance microprocessors, numerical analyses to investigate the concept of intentional transient power manipulations in microprocessor cores, transient thermal characterization of servers, and highly energy-efficient liquid cooling systems for data center facilities. In addition, we have our traditional favorites, which are the calculation corner and the thermal facts and fairy tales sections. I hope that you enjoy this stimulating issue of ElectronicsCooling and would like to offer you all my warmest wishes for a great holiday season. Stay warm or cool based on your preference, but I hope it is thermally energy efficient!
