Contents I. Introduction
With the 60 GHz IEEE 802.11ad technology, this paper proposes a novel medical imaging cloud network architecture that consists of (i) in-hospital medical image big-data database, (ii) an AP scheduler, and (iii) IEEE 802.11ad access points (APs) [1]. The medical image big-data database is connected to the all deployed IEEE 802.11ad APs via the AP scheduler. The deployed IEEE 802.11ad APs can be densely located within large-scale hospital building. In this densely deployed network topology, the closely and densely located IEEE 802.11ad APs can generate interference among them. This is the main reason why the proposed hospital healthcare cloud architecture requires an AP scheduler. The medical machines (including magnetic resonance imaging (MRI) and computed tomography (CT)) and users (including doctors and patients with wearable sensors) in the hospital networks can be wirelessly connected to the medical imaging big data database via the randomly and densely deployed IEEE 802.11ad APs. Moreover, the sizes of medical images are quite big, therefore, the queue management in each IEEE 802.11ad AP is important (to deal with potential queueing delays and overflows in each AP), and this management operation is closely related to the AP scheduling. Therefore, this paper proposes an AP scheduling algorithm for this proposed hospital healthcare medical imaging cloud architecture under the consideration of queue-backlog sizes. For the AP scheduling,
This paper formulates the problem with the theory of a max-weight independent set (MWIS) [2].
Based on the MWIS formulation, this paper makes a synchronous scheduling decision with discrete-time queue-weighted learning.
