Requirements and Challenges

Application overview

Regional Healthcare Information Network (RHIN) applications primarily include portal services, healthcare collaboration, public health, and research management. Public services, drug management, and other healthcare services can also be provided.

Portal services enable the public, doctors, healthcare administration personnel, and researchers to access public portals for browsing and obtaining requested information.

Healthcare collaboration refers to the inter-organization services within a local region, such as the referral between a community hospital and a specialist hospital, Electronic Health Record (EHR) query, and teleconsultation services.

Public health collects EHR information within the local region, identifies diseases, authenticates the spread of a disease, controls diseases, and manages chronic diseases.

Research management analyzes and taps statistical data for healthcare organization performance management and research promotion in the healthcare field.

System pressure

As service systems go online, the integrated RHIN system needs to process an increasing amount of data. One traditional way to solve the problem is to simply purchase more servers to cope. However, one or more new servers can support only one application, which deteriorates system performance, consumes excessive equipment room space, and wastes power supplies and cooling resources.

In a traditional system, hardware resources assigned to each application are configured based on performance requirements at peak hours. However, during off-peak hours, these hardware resources are unused, wasting energy and resources.

Scalability management

Against this backdrop, dynamic scalability management becomes a natural choice. For example, at the peak hours of portal access requests, more Web server instances and inter-cluster load balancing are provided to enable a large number of access requests.

Scalability management is designed for cloud applications. The cloud platform starts a lot of instances to handle huge amounts of traffic and stops some idle instances when traffic decreases. To this end, scalability management dispatches resources according to the actual needs of cloud applications.

Administrators configure application deployment templates in advance, including the application images, Virtual Machine (VM) specifications, topology connections, startup initialization, and parameter configuration scripts. Adding servers for application services enhances only resource computing capabilities. Cluster technologies are needed to integrate computing capabilities and provide services as a whole. When the application load reaches its maximum, the cloud platform performs scalability management based on preconfigured scalability policies.

(1)The Infrastructure as a Service layer creates VMs, loads application images, and generates new application instances.

(2)Meanwhile, administrators create a cluster to integrate these instances. Scientific management of application systems is also required to balance the load between multiple VMs within the cluster.

Inter-application resource multiplexing

Portal services are useful as examples. Portal traffic reaches its peak during the day, and the traffic of the residential health portal reaches its peak at night. If VMs are deployed based on the highest number of service requests on each portal, resources will be wasted during off-peak hours and customer investment will be increased. However, with resource multiplexing − based on cloud scalability − the cloud platform releases most of the instances normally assigned to the residential health portal and dispatches released VMs to the professional portal during the day. At night, the platform dispatches most of instances to the residential health portal.

Inter-application resource multiplexing is applied to an array of events, because in practical operation, it is impossible for different applications to reach their traffic peak at the same time. If each application applies or reserves resources based on service requirements at peak hours, a large number of resources are wasted. With the help of scalability management, the system can release resources assigned to Application B (whose service request amount is rather small) and dispatches these resources to Application A (that carries the largest traffic in the day). This significantly utilizes inter-application resource usage, reduces overall resource requirements, and reduces customers' Total Cost of Ownership (TCO).

Huawei Solution

Huawei's integrated RHIN system uses a cloud platform. With the cloud platform, Huawei's system provides modular installation and deployment of application software, enables users to install software with just a mouse-click, and supports dynamic scalability management and inter-application resource multiplexing. This platform also provides dynamic resource dispatching and power supply management. After expanding capacity, the platform reconfigures the VM resource pool based on service capacity requirements for load balancing. After shrinking capacity, the platform centrally deploys VMs and shuts down idle servers for reduced power consumption.

The following is the procedure for deploying applications on the cloud platform:

1. Develop RHIN applications.

2. Develop application deployment templates, including the components, connection and startup relationships between components, network topology relationships (such as port numbers, IP addresses, or internal domain names for the connections), VM specifications (CPU, memory, storage, and input/output bandwidth), and make images for application systems using intuitive GUIs.

3. With simple mouse clicks, administrators at RHIN data centers deploy the entire application system except for instance parameters (such as IP addresses for the external applications) based on specified service requests. The application system includes the modular components, databases, operating systems, and software for single-node or multi-node systems.

4. Administrators configure scalability management policies, including the number of VMs during the initial deployment and policies for expanding or shrinking capacity. These policies comprise resource thresholds and the number of increased or reduced VMs during every capacity change. 5. The system also enables application callback. Dedicated interfaces are provided to prepare the system for capacity changes such as load balancing after capacity expansion.

5. The service systems automatically operate based on specified capacity management policies. Administrators need to ensure the suitability of the policies by referring to the system operating status during policy configuration.

Scalability management policies provide the following benefits:

Huawei's cloud platform monitors applications' Key Performance Indicators (KPIs) in real time and counts monitored results using statistical methods to determine whether a fault occurs on the current application. If the application encounters a fault, monitored results can help rectify it.

Customer Benefits

Dynamic scalability management and inter-application resource multiplexing increase platform flexibility, enabling administrators to better cope with a large number of requests at peak hours and avoiding waste by holding resources in reserve.

Huawei's RHIN cloud platform deploys service templates in an intuitive manner and supports an array of advanced functions, such as inter-application resource multiplexing, KPI customization for flexible monitoring, and scripts and interfaces for service callback. These functions help simplify platform deployment and enhance platform performance.