Heterogeneous Computing in the Edge
Performance Ratio = Some throughput measure / Some cost measure .
In general , heterogenous computing in the edge seeks to increase the throughput values and / or reduce the cost values so the overall performance ratio of the system is optimized . Of course , optimizing these values for a certain application does not guarantee that it will be optimized for other applications . This implies that different configurations and complements of processor types may be required for different application sets , and no single configuration will serve all applications .
One important cost measure is the dollar cost to purchase the processor hardware . Getting more throughput per dollar of system purchase price is one important way of optimizing the total lifecycle cost of ownership of an edge system . The amount of throughput a dollar will buy is highly dependent upon system architecture , the capabilities of the processor chips , the efficiencies of the hardware and software infrastructure , and the requirements of the software and algorithms .
Another cost metric is energy used . Electrical power needed to continuously operate an edge node is usually the largest component of its ongoing operational expense . In many edge node applications , this power is supplied by batteries ( at least during the times when AC power is unavailable ), and adequate battery capacity for sustained operation is very costly . Also , the electrical energy that enters a processor chip is almost completely converted to heat that must be removed from the system , and the necessary cooling infrastructure is a strong contributor to the purchase and operational costs . Power and cooling can create absolute limits on the throughput of edge computers .
Space is another cost driver . Edge computers are located outside of traditional cloud data centers , in facilities like huts at the base of cell towers , roadside cabinets , underground vaults , micro data centers in shipping container-like enclosures , and mobile enclosures riding on vehicles or carried by humans . As processors get physically larger , the cost associated with providing that space grows rapidly . Often , it is impossible to support edge computer designs for a given network deployment model if their physical size exceeds a certain value , and this can constrain system throughput .
Weight is the final cost driver we will discuss . In certain deployment situations , especially aerospace , maritime or human portable deployments , there are strong constraints to the maximum weight of an edge node . The choices of processor technologies can have a strong influence on the overall weight of the system . If the weight of an edge node exceeds a certain limit , it may not meet the system requirements , and it will have to be redesigned or its functionality will have to be adjusted .
IIC Journal of Innovation - 47 -