WavebreakmediaMicro - Fotolia
Untangle cloud application integration spiderwebs
A migration is a big feat, but once you are in cloud it's not a 'set it and forget it' situation. Enterprises should take a closer look at integrations and connections.
When some organizations contemplate a shift to the cloud, they envision that their cloud provider will simply take over all their IT needs. Of course, it never really works out that way.
An enterprise can leave on-premises equipment and software management behind when it moves to the cloud, but it's still responsible for cloud application integration. But with so much focus on the resources freed up by a migration, the application's dependencies within the organization are often overlooked.
Today, few workloads are just a single application in a stand-alone environment. Systems such as electronic health records and email interact with each other to provide a feature-rich environment for internal staff, and users expect those tools to integrate so they can provide quick and complete services to their customers.
However, when there is poor cloud application integration and the connection breaks, it negatively affects the customer experience. So whether you're talking about IaaS or SaaS or anything in between, how your enterprise interfaces with those providers is essential. Once an application is in cloud, you must find all the integration points that could cause failure.
A majority of applications use APIs to communicate with each other and share data. APIs are critical links that enable seamless integration between systems. These links function because the software vendors have agreed on common data exchange standards to prevent potential communication issues.
However, APIs often change in form and function when a vendor adds features or makes other alterations, often for security purposes. These adjustments can hinder full and open communication between endpoints. With on-premises applications, enterprises can alter API security settings or even delay updates until after testing. But, this is not possible in the cloud.
A cloud provider may push users into a new set of API security standards, even if an individual user isn't ready to support it. That enterprise will have to scramble to ensure its applications support the update or risk losing API integrations. If the latter scenario occurs, the cloud application itself might not be offline, but it might as well be if its overall function doesn't work as expected. The enterprise will have to rush to fix it or find a workaround.
For example, this can happen when a vendor moves an application interface from a known software platform such as Java to something newer, like HTML5. These interfaces can break if they were written for Java, so the enterprise will need to update them. This isn't necessarily difficult if the updates are available, but it's still work that needs to be done -- most likely at a time that is not ideal.
While the application might be hosted in the cloud, authentication is often still tied to the data center. In this scenario, if the primary on-premises location goes offline -- due to a physical issue, bandwidth problems or a simple denial-of-service attack -- the cloud-based application will not be accessible to users, even if it continues to run. Without authentication, it's possible to lose access to any application that uses single sign-on. For example, it could cut users off from an application such as Office 365, which would render staff unable to do simple document tasks.
An enterprise needs to ensure the links and interfaces lead back to its data center so it can actually take advantage of the uptime benefits of the cloud. This doesn't mean users should avoid the cloud. The key is to not fall victim to the spiderweb of connections and understand how the cloud environment truly works within an application stack, as well as the effect it can have when problems arise.
Open the lines of communication
The first step to untangle potential cloud application integration problems is to map out all communications and dependencies. It can be difficult, if not impossible, to find the appropriate documentation for your own software and interfaces, but application mapping software, such as SolarWinds or VMware vRealize Network Insight, can help organizations identify the communication interactions. Mapping enables users to know who is talking to whom, at which point IT teams need to assess the effects of critical communication channel failures. Such mapping will enable enterprises to understand which communication channels are critical and what to expect when things do go down.
Of course, understanding the impact won't protect users from an outage. But if enterprises track the scale and details of the issue, they can notify users quicker and help mitigate communication problems. These plans can also help IT teams understand where they should deploy additional resources for redundancy and workload protection. A focused approach that reveals risks can go a long way when budgets and other resources are not always available.