I’ve been researching cloud applications lately, and based on my research, I am concerned that imaging services can quickly get confused by all the IT terminology mumbo jumbo! I thought I would take a stab at reviewing several options to raise awareness and make sure service providers understand where vendors are coming from.
Let’s start with the objective, namely disaster recovery. Healthcare facilities generate a lot of patient-sensitive data and healthcare regulations (such as HIPAA) require that facilities be able to produce that data to patients in a timely manner. If all data is kept on premises in one storage device, there is the risk of a catastrophic failure and loss of data. For that reason, many facilities duplicate the data and store in a secondary location in the event of a disaster with the primary storage.
In the past, facilities have for all intents and purposes been on their own with respect to managing the data replication and storage from both cost and infrastructure perspectives. The advent of cloud storage and computing is opening new doors in terms of how data can be managed for disaster recovery purposes. There are quite a few options that take advantage of the cloud as part of a disaster recovery strategy, and knowing the difference in options is important.
Backup and Restore from the Cloud
Think of this alternative as the simplest use of the cloud, as it simply becomes the location of a secondary backup, in lieu of a local backup media. Imaging applications and the data remain on site, and the cloud is simply the location for the backup copy. In the event of a loss, data is simply restored from the cloud to the local media. The major drawback is data restoration within the expectations of users. The earliest applications supported this model and were simply remote backups.
Backup To and Restore To the Cloud
In this case a data restoration is not made directly to the site but instead within the cloud, such as to a virtual machine. An example of this approach might be the addition of a cloud-based viewing application which would mean that information could be accessed through the cloud without full restoration to the local site. This might provide the site with some degree of business continuity, and provide users with a better expectation in terms of data access.
Replication within the Cloud
As replication implies, both the data and application are replicated in the cloud, thereby providing faster recovery in high-demand situations. Besides storage, this alternative requires replication of applications in the cloud in a virtual environment. Obviously, the application would need to lend itself to virtualization and remote operation, which may be limiting factors.
Managed Service Provider
In this instance, the application and data are managed by a 3rd party in the cloud, in terms of both the primary and disaster recovery instances. In tradeoff with reducing the on-site support of the application, the facility becomes totally dependent on the 3rd party for meeting both recovery time and security requirements – a key consideration when it comes to Patient Health Information (PHI)! Under the right circumstances, this approach can be attractive in terms of relying on the 3rd party’s larger infrastructure and knowhow, but it is also dependent on the 3rd party’s ability to meet the site’s performance expectations.
Meeting Expectations
I have been involved in situations where a vendor may be speaking to one alternative and the facility another! It is important to fully understand the capabilities of the vendor before making a commitment. For example, suppose a site really needs an ability to assure user access to the data in the event of a disaster. The simple solution of backing up and restoring from the cloud will not necessarily achieve this within user expectations. Using applications in the cloud may be more appropriate if the facility’s staff needs image accessibility even during an outage of the local application.
That is why it is important to have a clear understanding of application objectives including Recovery Time Objectives (RTO) and Recovery Point Objectives (RPO), analysis of local versus cloud-based costs, documentation of user expectations, the prospect for application vendor changes, and potential future applications such as Health Information Exchange (HIE) participation.
A good foundation in understanding vendor offerings, and a clear strategy for image disaster recovery are tantamount to consideration of cloud-based alternatives. I hope this dissertation is helpful to those contemplating an enterprise image management solution.
