As your company moves further into cloud computing, you may grow to appreciate the benefits of allowing third-party...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
providers to manage hardware infrastructures. But that’s exactly when you’ll realize there are other dependencies, workflows and architectural considerations you need to track in-house.
The management tools and procedures you developed for on-premises, dedicated and virtualized servers can meet some of your needs in the cloud, especially if your cloud-based applications haven’t significantly changed. But once you start taking advantage of new cloud capabilities, you’ll need to find new ways to manage cloud systems and functionalities.
For example, if your business collects and generates large volumes of data, running hundreds of servers means you can easily mine large data sets in a cloud cluster. You provision the cluster only when needed and then shut it down as soon as data-mining jobs have completed. But when it’s time to start another data-mining project, can you be sure all necessary images will still be in the service catalog? Will you remember how to deploy and properly configure databases, analysis servers, workflow engines and load balancers?
When you’re running complex configurations or using multiple cloud products, it’s time to turn to a cloud management service.
Additionally, your organization probably requires defined levels of performance and availability from mission-critical applications. One way to ensure you meet these requirements is to replicate servers in separate service regions within a single cloud provider or across multiple cloud providers. Well-publicized outages at Amazon and Microsoft highlight the potential risk of using a single cloud provider; however, you might be willing to take that risk instead of spending additional resources to develop and maintain separate scripts and manage resources in different clouds.
Both of these examples highlight the need to manage distributed but highly integrated systems. Self-service portals can be used to configure multi-component systems, but manual configurations are prone to error and inconsistencies. Many cloud players often offer services that support common configuration tasks, such as adding servers to a load-balanced group of servers.
As your cloud management needs become more complicated, you may find yourself falling back to using manual configurations. And since there’s no single industry standard management API for cloud providers, you may have to maintain multiple, provider-specific procedures when you use more than one cloud services provider. When you’re running complex configurations or using multiple cloud products, it’s time to turn to a cloud management service.
Cloud management services provide management functions that complement what vendors offer, including cloud configuration and deployment services as well as cross-platform capabilities. For example, a cloud management service might offer multi-component templates that allow you to deploy n-tier applications as a single unit using pre-defined templates. These services can also provide consolidated reporting and monitoring, which can be especially important for businesses in which multiple users have authority to deploy instances in public clouds.
Cloud management services can also help control system administration costs with event-driven alerts and configuration changes. System administrators can specify performance criteria and alerting thresholds for management services for incidents that may require attention, such as adding or removing servers from a cluster as demand changes.
Factors in choosing a cloud management service
There are several factors to consider when evaluating a cloud management service. Choose a vendor that supports your current cloud providers or those you intend to use. Smaller Infrastructure as a Service (IaaS) and Platform as a Service (PaaS) vendors may have suitable options, but the fact that they’re not supported by cloud management services makes them less viable for enterprise use.
Evaluate support for identity management, especially across multiple cloud providers. Do providers have sufficient access controls? Do they work across various cloud providers? Does the cloud management service offer key management functions? Identity management has always been challenging; with self-service portals, users can generate and revoke keys, further complicating key security and tracking.
It’s also important to understand how the service supports core system administration functions, such as cloud backups and reporting. Can you automate backups based on backup policies? Is reporting sufficient for your enterprise’s compliance and management needs? If you use chargeback or showback to report on cloud usage, look for a service that provides consolidated reporting divided by end user or business unit.
Finally, comparing prices among cloud management providers is not a trivial piece of the assessment. Some vendors charge by how many components are managed per hour while others offer tiered monthly pricing options. Consider typical use cases for cloud computing jobs and how frequently these jobs run. If you need to compare different pricing models, consider a combination of low-demand, likely-demand and high-demand scenarios. You won’t necessarily be able to precisely predict what your overall costs will be, but you will get a sense of the variation you could see in your monthly bill.
Dan Sullivan, M.Sc., is an author, systems architect and consultant with over 20 years of IT experience with engagements in advanced analytics, systems architecture, database design, enterprise security and business intelligence. He has worked in a broad range of industries, including financial services, manufacturing, pharmaceuticals, software development, government, retail and education, among others. Dan has written extensively about topics ranging from data warehousing, cloud computing and advanced analytics to security management, collaboration, and text mining.