Universidade do Vale do Rio dos Sinos, Av. Unisinos, 950, São Leopoldo, RS, Brazil.
Comput Methods Programs Biomed. 2020 Jul;191:105403. doi: 10.1016/j.cmpb.2020.105403. Epub 2020 Feb 20.
Multiple medical specialties rely on image data, typically following the Digital Imaging and Communications in Medicine (DICOM) ISO 12052 standard, to support diagnosis through telemedicine. Remote analysis by different physicians requires the same image to be transmitted simultaneously to different destinations in real-time. This scenario poses a need for a large number of resources to store and transmit DICOM images in real-time, which has been explored using some cloud-based solutions. However, these solutions lack strategies to improve the performance through the cloud elasticity feature. In this context, this article proposes a cloud-based publish/subscribe (PubSub) model, called PS2DICOM, which employs multilevel resource elasticity to improve the performance of DICOM data transmissions.
A prototype is implemented to evaluate PS2DICOM. A PubSub communication model is adopted, considering the coexistence of two classes of users: (i) image data producers (publishers); and (ii) image data consumers (subscribers). PS2DICOM employs a cloud infrastructure to guarantee service availability and performance through resource elasticity in two levels of the cloud: (i) brokers and (ii) data storage. In addition, images are compressed prior to the transmission to reduce the demand for network resources using one of three different algorithms: (i) DEFLATE, (ii) LZMA, and (iii) BZIP2. PS2DICOM employs dynamic data compression levels at the client side to improve network performance according to the current available network throughput.
Results indicate that PS2DICOM can improve transmission quality, storage capabilities, querying, and retrieving of DICOM images. The general efficiency gain is approximately 35% in data sending and receiving operations. This gain is resultant from the two levels of elasticity, allowing resources to be scaled up or down automatically in a transparent manner.
The contributions of PS2DICOM are twofold: (i) multilevel cloud elasticity to adapt the computing resources on demand; (ii) adaptive data compression to meet the network quality and optimize data transmission. Results suggest that the use of compression in medical image data using PS2DICOM can improve the transmission efficiency, allowing the team of specialists to communicate in real-time, even when they are geographically distant.
多个医学专业依赖于图像数据，通常遵循数字成像和通信医学（DICOM）ISO 12052 标准，通过远程医疗支持诊断。不同医生的远程分析需要实时将同一图像同时传输到不同的目的地。这种情况需要大量资源来实时存储和传输 DICOM 图像，一些基于云的解决方案已经对此进行了探索。然而，这些解决方案缺乏通过云弹性功能来提高性能的策略。在这种情况下，本文提出了一种基于云的发布/订阅（PubSub）模型，称为 PS2DICOM，它采用多级资源弹性来提高 DICOM 数据传输的性能。
实现了一个原型来评估 PS2DICOM。采用了 PubSub 通信模型，考虑到两类用户的共存：（i）图像数据生产者（发布者）；和（ii）图像数据消费者（订阅者）。PS2DICOM 使用云基础设施通过云的两个级别中的资源弹性来保证服务可用性和性能：（i）代理和（ii）数据存储。此外，在传输之前对图像进行压缩，以使用三种不同算法之一减少对网络资源的需求：（i）DEFLATE，（ii）LZMA 和（iii）BZIP2。PS2DICOM 在客户端采用动态数据压缩级别，根据当前可用网络吞吐量提高网络性能。
结果表明，PS2DICOM 可以提高 DICOM 图像的传输质量、存储能力、查询和检索。在数据发送和接收操作中，总体效率提高约 35%。这种增益是由于两个弹性级别，允许资源自动向上或向下扩展，以透明的方式。
PS2DICOM 的贡献有两个方面：（i）多级云弹性，以按需适应计算资源；（ii）自适应数据压缩，以满足网络质量并优化数据传输。结果表明，在使用 PS2DICOM 对医学图像数据进行压缩时，可以提高传输效率，使专家团队即使在地理位置上相距很远的情况下也能实时通信。
