everythingpossible - Fotolia
These days a software developer can easily download a copy of a PaaS offering, run it and play with it in the sandbox. That's true whether the developer chooses AWS Electric Beanstalk, Google App Engine, IBM Bluemix, Microsoft Azure or any of the many other PaaS platforms available now. What's not easy is to have to support for your PaaS environment in production.
In answering this question, I can speak for IT managers, because I have been one. Nothing is more important than keeping your production systems up and running. One of the reasons many IT managers are cautious about using PaaS platforms or public cloud is loss of control over keeping their IT infrastructure up and running. While you may own the responsibility when they come knocking at your door, you don't own the infrastructure or the support of it and you simply cannot control the cloud provider's maintenance procedures.
That is where either a service-level agreement (SLA) or service-level objective (SLO) comes into play. The SLA must define the relationship and set clear expectations for what the service levels are between you (the buyer) and the cloud provider (the seller). At a minimum, you will need to understand how the PaaS platform's cloud SLA defines production support.
Let's summarize some of the more popular PaaS platforms and what they offer from an SLA perspective and resilience or availability perspective.
Azure has been in production for over eight years now, which gives it the product maturity that IT managers crave. Stipulated uptime is 99.95% for standard-tier customers for this PaaS platform; but that guarantee doesn't extend to free or shared tiers. There is high availability and redundancy in data centers globally. There is also 24/7 tech support and health monitoring. If your uptime is 99%, you get a 25% service credit.
Azure has a global presence and the storage of data can be restricted to a single region. That is important to customers who need to identify where there data resides, because of internal contracts you might have that preclude the data from residing outside a specific region. Azure even has a service called Azure Site Recovery, which provides a mechanism to failover an on-premises data center. When disaster strikes, there is nothing more important than having a way of recovering your infrastructure.
IBM's Blumix cloud infrastructure is hosted and managed within IBM's SoftLayer data center, which provides for hardware and network redundancy and is also monitored by IBM staff 24/7. IBM offers a 99.95% availability service level for service instances and runtime instances that are configured for high availability across two Bluemix regions. IBM also provides the ability to have a private cloud, even allowing you to host your infrastructure in your own data center. This is important for companies that want to use some of the advantages of cloud computing but simply will not allow for their infrastructure to be hosted off-site. No other vendor provides this type of cloud flexibility and the myriad of options that IBM can offer.
AWS Elastic Beanstalk
Amazon does not provide any guarantees specific to Elastic Beanstalk, though it does provide an SLA for their Elastic Compute Cloud -- EC2, which is a central piece of Amazon's cloud computing platform. EC2 is one of Amazon's integral services which provide access to the server instances. In other words, while AWS is a set of services from Amazon's overall laaS offering, EC2 is perhaps their most important service. If this service is available less than 99%, you will see a 30% service credit.
Google App Engine
Google App Engine provides an SLO of 99.95%. If Google does not meet this objective and the customer meets their obligations, the customer will be eligible to receive financial credits. If it's less than 95%, you can receive a 50% credit. Between 95% and 99%, you would receive a 25% credit. Similar to the other offerings discussed, Google provides multiple data centers across the globe, to ensure redundancy and availability.
While each of the vendors offers some kind of SLA, you will need to dig under the covers to find out more about what they offer from a resilience and high availability standpoint. While each of the vendors put their infrastructure into secured redundant data centers, make sure you ask the specifics about their high availability locations and also if they can provide penalties to you, in case of certain outages -- not just usage credits. If your business is large enough and important enough to them, they might even tailor their existing SLO or SLA contracts around your unique requirements.
A close look at Google App Engine
Pros and cons of three PaaS offerings
App modernization starts with PaaS
Dig Deeper on PaaS and other cloud development platforms
Related Q&A from Kenneth Milberg
Learn which makes more sense for your enterprise: building your own cloud-based diagnostic tooling or purchasing an existing product. Expert Ken ... Continue Reading
Learn about Helm Charts and how Kubernetes supports a microservices architecture. Expert Ken Milberg discusses what you need to know before delving ... Continue Reading
As blockchain matures, an ecosystem of tools continues to sprout up around it. Learn how these offerings, including the Hyperledger Fabric Client SDK... Continue Reading