ra2 studio - Fotolia
Offering the flexibility to use a mix of public and private infrastructure, hybrid cloud has become a common goal for many enterprise IT teams. But the implementation of that model can introduce significant challenges.
From proof-of-concept (POC) projects to orchestration best practices, follow these five helpful tips to smooth the road to hybrid cloud deployment.
1. Perform a thorough hybrid cloud POC
An effective hybrid cloud strategy starts with a thorough POC. Take four specific steps to ensure a POC project covers all the bases.
First, evaluate how hardware changes will affect the migration of existing VMs to the public cloud. Chances are that public cloud providers run their infrastructure on newer, more advanced hardware than what an enterprise has on premises. Prepare for differences around memory, solid-state drive performance and network bandwidth. Try a sandbox testing environment to choose and optimize public cloud compute instances.
Next comes data migration. Delete unnecessary files, and determine where to host each data set -- strictly on premises, in the public cloud or across both. To perform the actual migration, use either the piecemeal method, which occurs in phases and often without downtime, or the big bang approach, which requires downtime.
The third step is to set up cloud bursting and networking for your POC. Due to the lack of cross-cloud standards, building scripts to set up virtual networks in the public cloud can be complicated. Use a translation tool to make the transition smoother.
The final step is resiliency and failover testing. Run through each process multiple times to ensure that it functions properly and that operations can continue in the event of a disaster.
2. Prepare apps for hybrid cloud deployment
A successful hybrid cloud implementation starts with the division of application components into the hosting environment -- either public or private cloud -- in which they'll best run, based on cost, performance and other factors.
As many IT teams choose to host application front ends in public cloud and application back ends in private cloud, the next step is to design the communication workflow for APIs across those two environments.
Then, control a virtual private network (VPN) address space by dynamically assigning IP subnets to users and applications. For public cloud addresses, learn how your cloud provider handles IP addresses, and map them consistently into your VPN and directory spaces.
Finally, have a load balancer that supports components in both the cloud and the data center, and perform tests to ensure it operates effectively in hybrid cloud.
3. Evolve security practices for a hybrid cloud strategy
Don't let security be an afterthought in a hybrid cloud deployment. Rather, establish solid security plans in your initial design to avoid lapses in protection.
First, understand the shared responsibility model of your cloud provider. Maintain an open line of communication to know where exactly the provider's security responsibilities end and your own begin.
4. Master hybrid cloud orchestration
Hybrid cloud requires an effective orchestration platform to maintain consistency across environments. When coupled with automation, orchestration also eliminates manual IT provisioning, bolsters security and minimizes troubleshooting.
In a hybrid cloud deployment, an orchestration tool must be able to determine whether a VM instance would run more efficiently -- and cost-effectively -- in public or private cloud. In many cases, organizations will need to adopt a third-party orchestration tool for hybrid cloud, as public cloud providers' native orchestration tools won't always work outside their respective platforms.
5. Improve hybrid cloud monitoring through automation, alerts
To maintain a successful hybrid cloud deployment, know how to monitor it. Certain best practices can help.
Collect a subset of monitoring data from the various cloud platforms you use into a centralized monitoring tier. This standardizes monitoring practices and ensures alerts are sent to the right people.
Next, define important key performance indicators to monitor, and continuously re-evaluate them to prioritize those more critical to the business. Also, establish a balanced alert threshold. Too many alerts cause alert fatigue, while too few fail to properly flag issues.
Connect monitoring tools with communication apps, such as Slack or PagerDuty. This way, you can see how often the IT team addresses a specific issue. When that same issue occurs again, use prior communications to more quickly resolve it.