Who's Managing Your Cloud?

The expectation that cloud providers will manage enterprise systems and applications is sometimes misplaced. Be sure to create effective management guidelines from the start.

Tim Brophy

October 15, 2015

5 Min Read
NetworkComputing logo in a gray background | NetworkComputing

IT leaders in many organizations are choosing to outsource their operational tasks to specialist service providers. When this happens, in-house staff can ideally be freed up to perform more proactive roles, rather than spending valuable time keeping up with administrative and routine tasks.

As more and more applications are offloaded to cloud service providers (CSP), outsource agreements must also take into account the different delivery models of cloud services. The typical use case for cloud hosting has been in non-core environments such as dev, test and user acceptance testing (UAT), which have been largely unmanaged. However, with cloud gaining credibility as a platform to host production and disaster recovery workloads, companies will need additional structure around these applications and their management.

But isn't the cloud already managed?

By default, the cloud is managed in the sense that the underlying services infrastructure is abstracted from the client and managed by the CSP. This is true in varying degrees for any of the cloud environments, whether it is Infrastructure-as-a-Service (IaaS), Platform-as-a-Service (PaaS) or Software-as-a-Service (SaaS).

The greatest level of management by the provider is delivered in a SaaS environment, in which the customer need only concern themselves with the configuration and data customization of the software service. The provider ensures that the entire application is available and operational.

This is similar to a PaaS environment, but the primary difference is that the client owns and is responsible for the data, codebase and configuration of the application. The execution platform and OS is delivered by the provider following the agreed upon service level agreement (SLA). 

IaaS, however, provides for a different dynamic. In this instance, the cloud provider provisions the operating infrastructure upon which virtual servers are built and operating systems installed. The IaaS provider by default does not manage the operating system nor the application -- this responsibility fall upon the client.

Given that a large percentage of daily operations pertain to OS management and related activities (such as capacity and performance management) this means a significant amount of maintenance is still needed on the customer end. As a consequence, many organizations that move to the cloud in the hope of offloading mundane administrative duties succeed in respect of hardware management only, while still being tasked with the need to manage operating systems, etc. 

Cloud management vs. on-premises

So clearly there is a need and place for managed service providers within the cloud services landscape, but the process of choosing providers and implementing agreements needs to take key differences between traditional and cloud environments into account.

First, cloud deployments tend to be dynamic -- not static -- and must grow or shrink depending on the requirements of the client. Any agreement with a managed service provider (MSP) needs to cater to the variability that the consumptive nature of the cloud provides.

Secondly, IT leaders must consider who has access to the provisioning of resources. Additional control may be required by the provider to ensure that they can meet uptime obligations, but processes must be sufficiently streamlined to avoid nullifying any of the agility that cloud platforms bring with them. By this same token, the IT customer should rest assured that their costs are not spiralling out of control through actions taken on their behalf by the MSP. 

Depending on the outsourcing strategy of the company, either one or multiple MSPs may be chosen to deliver services based on focus area. A single vendor strategy may result in an ease of engagement and interaction, but often their areas of expertise do not overlap to all of their client's requirements. Even if the vendor has an area of competence across all areas, these capabilities need to be vetted to ensure that they can be sustainably delivered.

For example, a client may want to leverage an MSP for database work on their Oracle systems, but their chosen MSP may have limited resources and skills to manage that environment. In such a case, it would be far more advisable to leverage a specialist Oracle MSP, while continuing to use the existing MSP for their broader workloads and functions. 

When a provider must manage a blended or hybrid environment, various intricacies will arise regarding the integration of multiple systems and platforms. Monitoring tools that the MSP currently uses may not be cloud-ready, and require access to parts of the platform, such as the hypervisor, which the CSP simply cannot expose. Backup tools and strategies may need to change, and disaster recovery plans will need to factor in the different methods of replication and failover that can be achieved. 

Many providers may be prescriptive about the platforms their clients leverage as part of the agreement due to the MSP's familiarity with certain cloud platforms. Ultimately the MSP is delivering an outcome and needs to be sure that the platform and the associated tools and capabilities can support this outcome. The accompanying service architecture needs to ensure that responsibility for specific service areas can be correctly allocated at all times to avoid finger pointing. In the case of application performance, many different factors need to be taken into account. The accountability to maintain application performance and find root causes to incidents must lie with a single party in order to keep multiple vendors from blaming each other.

In all instances, IT leaders must be clear on their requirements, not only for the benefit of providers who need to tailor their solution, but also so an adequate level of expectation can be set on the MSP and CSP by all parties that make up the agreement. Applying additional levels of service to those traditionally offered by CSPs is a necessary step in the lifecycle of an enterprise's adoption of cloud as an IT service delivery model. It enables them to accelerate their journey by leveraging the skills and expertise of MSPs with vast experience in managing not only CSPs but also operating environments efficiently. 

About the Author

Tim Brophy

Global Cloud Architect, Dimension Data

SUBSCRIBE TO OUR NEWSLETTER
Stay informed! Sign up to get expert advice and insight delivered direct to your inbox

You May Also Like


More Insights