1. Introduction 2
1.1 Purpose for this Paper 2
1.2 Target Audience 2
1.3 Provide Primer to Cloud and Cloud Terminology 2
2. Discuss Relevant HL7 Standards (“the what”) 2
3. Present the Cloud Blueprint(s) 3
3.1 Presenting the Blueprint Concept 3
3.2 Architectural Considerations 3
3.2.1 Deployment and Migration Considerations 3
3.2.2 Risks 3
3.3 Blueprints (each of which will include HL7 standards applied to the respective cloud blueprint) 3
3.3.1 Blueprint #1 (single institution, private cloud) 3
3.3.2 Blueprint #2 (hybrid environment, federated) 4
3.3.3 Blueprint #3 … 4
4. Security and Privacy 4
5. Cloud Maturity Model 4
5.1 Maturity Model for Healthcare Cloud Implementation 4
5.2 Organizational Cloud Readiness Self-Assessment 4
6. Potential Futures – “The Art of the Possible” 4
7. APPENDIX I: GLOSSARY 4
8. APPENDIX II: Resources and Recommended Reading 4
1. Introduction
1.1 Purpose for this Paper (Ken w/Diana supporting)
The HL7 Cloud Planning Guide was conceived to establish a quality set of references, best-practices, and considerations relating to the use of Cloud as applied to health organizations. Cloud is a significant information technology (IT) movement yet has little or no presence within HL7. That said, based upon an open survey of HL7 members, there was strong indication that many were undertaking cloud-based efforts, and often as part of modernization strategies.
The Hl7 Cloud Planning Guide is intended to bring together a collection of best-practices, repeatable patterns, common considerations, and healthcare-specific requirements that health organizations might consider as they undertake their own cloud efforts. For organizations amidst cloud efforts, this can serve as a validation of plans or as a legitimator of approaches underway. For organizations in early efforts, the document can serve as a skeleton to help in cloud planning. The following “FAQs” are based upon questions that came up repeatedly during the development of this document:
Cloud is about re-platforming. Why is HL7 doing a cloud paper at all? There was an underlying hypothesis, validated by an industry survey, that interoperability and/or HIT modernization were driving factors influencing cloud plans among a large constituency of HL7 members. The emphasis of this work is to highlight relevant healthcare-specific considerations, best-practices, and applied use of HL7 standards where used to further objectives as part of cloud implementation.
Aren’t there a ton of papers already out there? Why re-create one? There is no intention to create “yet another” cloud white paper, especially where there are many very good sources out there. That said, one of the goals of this work is to provide a “primer” to those less familiar with key cloud concepts, and to serve as a reference, drawing the reader’s attention to what have been determined to be high-quality reference sources (from consortia, academia, government, and industry/vendors). None of the included references should be considered endorsements of specific product offerings or platforms.
Is this an ‘implementation guide’? No. Other organizations are much better suited to provide cloud implementation guidance. This aims to be at a higher conceptual level, focusing instead on key influences, requirements, processes, and best practices to inform your implementation plans.
Doesn’t the notion of healthcare-specific requirements negate the value of Cloud? Not necessarily. While one of the true benefits of cloud is in information ubiquity and on-demand capability, elements of the healthcare vertical place constraints on what is or is not acceptable. This paper aims to raise awareness of some of the issues that have affected early adopters with the intent of benefitting readers. For instance, “classic” cloud allows data to reside anywhere, providing location independence, but many countries and localities require health data to be stored within geographic boundaries (e.g., within a country or territory). While limiting, this is a real, practical constraint on how and in what contexts Cloud can be used, or adaptations that must be considered to apply Cloud to health situations.
So what do HL7 standards have to do with this? Migration to cloud is often accompanied with a shift in IT service delivery models. This is not always the case (e.g., replatforming of an EHR into the Cloud is just that), but when that migration also includes modernization activities, such as web-enabling, mobile-device enabling, or service-orienting an offering, interoperability comes into play. Existing and emerging Hl7 standards have a potential benefit in these scenarios, which are a focus of this work.
Reading this Document. This document is organized into three principal sections: Blueprints, Security and Privacy, and the Cloud Maturity Model.
The Blueprints Section is a collection of best-practices that have been mined from user experiences and contextualized into business situations. Following the ideas of the classic computer-science reference text by Erich Gamma and the “Gang of Four”, this section presents design patterns in context, characterizing problem classes and then presenting potential approaches based upon that situational awareness.
The Security and Privacy Section focuses on special considerations to be taken to address these concerns and sensitivities for health and protected data. Note that this section does include realm-specific examples as illustrations, but should have equal applicability across the needs of many situations, countries, and regions.
The Cloud Maturity Model provides a basis for the objective evaluation and ultimately improved delivery capacity relative to cloud plans and execution. This is intended to serve as a catalyst both for identifying areas of potential improvement within your organization, as well as charting a course for capability sets recognized to improve efficacy and ability to execute cloud activities. Ultimately, this serves as a backdrop to help organizations identify and remediate potential exposure areas to a more effective outcome and delivery.
1.2 Target Audience (Vadim)
As cloud technologies have emerged and continue to develop they are impacting technical and business infrastructure. Several drivers are causing organization to refine their approaches to how systems are built and managed:
- Cost
- Revenue
- Convenient and rapid acquisition of resources
- Access to vendor ecosystem
- Security
- System and Solution Architecture (Could Native Apps)
These drivers impact business and technology strategy of interest to various stakeholders in various roles:
CTO/CIO is responsible for long term technology strategy of the organization and will need to consider appropriate adoption of cloud technology
Enterprise Architect - responsible for ensuring alignment of business and technology strategy and will have to identify appropriate technology, including the cloud, to support business strategy
Chief Security Officer is responsible for overall security of organization’s assets including physical, personnel and information assets and cloud technology and will need to consider risk and compliance of cloud technology adoption.
Business Owner is responsible for efficient delivery of services and products to the customer. Cloud technology allows rapid provisioning and scaling of infrastructure and applications to address transient or experimental business needs.
Other thoughts:
- What security standards could be applied to the Healthcare cloud?
- Are there emerging services that make cloud even more useful in this healthcare? Concerns that make it less useful.
Typical patterns of use include:
- Lift and shift
- Cloud bursting
- Analytics
- Hosting de-identified, aggregated, population data
1.3 Provide Primer to Cloud and Cloud Terminology (Vadim)
At its core, cloud is defined by two primary characteristics:
- Cloud is an economic model for acquisition of resources where resources and metered and accounted for based on some measure of utilization, e.g. duration the resource was available for access.
- Cloud offers rapid elasticity though automation and self-service. The consumer can rapidly acquire and release resources without operator intervention. This characteristic distinguishes cloud technology from virtualized infrastructure environments.
By offering these two primary characteristics, IT organization fundamentally change both economic and operational processes for technology use and adoption. For example, by adoption cloud technology the enterprise gains the ability to rapidly deploy infrastructure for transient or experimental business and technology efforts without capital investment.
At the foundation of cloud technology is virtualization of infrastructure assets, such as servers, storage and networking. Virtual servers, software defined storage(SDS), and software defined networking(SDN) decouple infrastructure available to the consumer from physical hardware. Server virtualization, for example, allows multiple virtual machines to run on single physical machine. Since typical physical server is utilized at a fraction of its capacity and is often idle, several virtual servers can be run concurrently on a physical server without significant impact to performance. This offers significant economic benefits. From the perspective of the consumer the virtual machine operates in manner that is indistinguishable from a physical server.
Cloud technology combines virtualization with management tools, automation and orchestration to deliver infrastructure rapidly and on-demand. Core characteristics of the cloud include:
- Shared pool of physical resources
- Sufficient scale to allow that resource pool to appear infinite for typical applications
- Ability provision or extend virtual resources quickly and automatically in response to changing conditions, e.g. application load and performance
- Services are metered and customers are charged based on utilization
- On-demand self-service capability to manage resources without staff interference
The three types of cloud IT service delivery are:
- Infrastructure as a Service (IaaS) delivers server, storage, and networking services typically available a data center. User is able to request resources, often using a self-service portal, and have those delivered in minutes. While the cloud provider manages the physical infrastructure the customer is able administer and access the virtual infrastructure.
- Platform as a Service (PaaS) delivers database, queuing, emailing, functional web services that a developer may need to construct an application or end-user service. The customer is able to administer setting or properties without the parameters allowed by the service, but is does not have access to underlying physical or virtual infrastructure.
- Software as a Service (SaaS) delivers software to the end-user. The application is typically running in shared infrastructure with users granted access to perform end-user functions on their data. The user has access to the user interface of the application and can only perform function exposed through that interface.
Cloud technology is area of active research and development with additional models of cloud computing emerging. New services, often referred to as “serverless computing” or and Function as a Service (FaaS), offer pure compute as a resource and allow deployment code without infrastructure to host the software components. Code is executed when triggered by an event, for example message arrival or a scheduled event. Resources required to process the request are allocated at the time of the event and only for the duration needed to complete the operation. User is only charged for the resources, CPU and memory, utilized to process the request.
Cloud technology deployment models include public or commercial, private or on-premises, and hybrid:
Public or commercial cloud offers the capability to acquire access to resources without acquiring underlying physical assets. The consumer relies on the vendor to operate the infrastructure, including physical security, networking, and data center operations, while allowing the consumer to provision virtual servers, storage, networks, and other infrastructure resources.
In a private cloud deployment, the organization owns the physical assets and makes them available for exclusive use of the organization. One can think of economic aspects inherent in the cloud in terms of inter-departmental accounting.
Hybrid cloud seeks to use both private and private cloud capabilities. Use cases for hybrid cloud include “cloud-bursting” where public cloud is used to handle occasional peak loads, and partitioning different workloads to private and public clouds. For example, private cloud could be used to process sensitive data and public cloud for public or less sensitive data.
Cloud imposes several considerations beyond purely economic concerns:
- Data management:
o In public and hybrid deployment models data leaves the enterprise. Security and data management practices of both the cloud vendor and the customer must be adequate to protect sensitive data in a shared security model. Strong cryptographic protection of data in-motion and at-rest is often a requirement for sensitive data and key management tools and practices becomes a core enterprise prerequisite for using the cloud.
o Cloud infrastructure controls physical location of the data and data location may not be readily apparent to the customer. Regulatory and contractual obligations may restrict permitted geographic location of the data.
o Data movement across physical infrastructure may require multiple copies to exist during transition. Technical and business processes must exist to track sensitive data location, and the customer must be able to locate and audit all copies of their data.
o Termination of services imposes requirements on exporting and purging of data in the cloud infrastructure.
- Security:
o Security is shared between the vendor and the consumer. In case of commercial cloud, the vendor is responsible for physical access to computing infrastructure. Coordination, review and auditing processes are required to ensure adequate security of the entire system.
2. Discuss Relevant HL7 Standards (“the what”) (Deferred until survey result received)
· To identify from within the HL7 portfolio of standards those items with relevance and applicability to a cloud environment.
· Identify relevant HL7 standards (HL7 SOA standards, related FHIR standards,…)
Ø Substantiate the viability and utility of HL7 SOA services in this setting
Ø Relate and contrast how FHIR fits within this environment
3. Present the Cloud Blueprint(s)
3.1 Presenting the Blueprint Concept (Ken)
· This project will define a "blueprint" to illustrate candidate approaches and fitness-for-purpose of HL7 assets into cloud environments, targeting HL7 technology adopters and implementers and providing guidance.
· Scope is to illustrate how cloud solutions can be design with a service-oriented approach. This is different from "classic" HL7 implementation, which is typically more about interoperability between organizations. Cloud is a bit different - the issue is how to design cloud-based solutions for health. Service-orientation is a natural marriage for this, but requires a different perspective on the problem space.