According to Datacom customer research presented in Before You Go Public, Read This, few organisations are currently planning to go ‘all in’ to public cloud and will, therefore, retain some of their workloads on-premise or in private cloud. Many services are, and will be for the foreseeable future, delivered via applications or workloads with a hybrid set up. For example, an organisation may want to use Microsoft Azure for a front-end application that needs elasticity, and use a private cloud for the interconnected database.
Managing hybrid cloud complexity
Hybrid cloud architecture, however, requires careful management and planning to account for key factors such as latency, security and compliance, as well as potential added complexity and transition costs.
“Making hybrid cloud work well means focusing on integration and interconnectedness, especially in the planning stages of a project. For instance, if an organisation stretches certain components by running them in Azure, what is the impact on other, reliant components? These critical factors are sometimes overlooked,” says Brett Alexander, Solution Architect at Datacom.
On top of this, with greater adoption of public cloud (and proper planning), usually comes a corresponding service orientation and increasing focus on business services and related outcomes enabled by cloud. These outcomes may include risk or reputation management, reducing cost and making key services available when needed and at a suitable quality. They are often delivered through the aggregation of multiple providers, services and solutions – and various SLAs.
Making these services and outcomes happen and managing the many moving parts involved is clearly an important function and a complex task, which organisations can take on themselves or outsource, at least in part, to a qualified partner like Datacom. Whatever the approach taken, there is a growing need for those in the organisation involved in Azure to understand the way different clouds and related services interact and how they can be integrated – with each other and with other environments and types of IT.
Monitor from business service level down
Business services, even something as simple as email, are built from and rely on a number of components, including applications, firewalls, switches, servers and storage. Mapping these services, the applications and infrastructure that enable them, and the interconnections and dependencies of the various components, are an integral part of planning for Azure adoption and optimisation and managing a hybrid cloud environment.
This is why Datacom recommends monitoring at an availability-of-business-service level. This means having dependency-based monitoring from the business service level down through applications and infrastructure, including Azure. We also recommend automated root-cause analysis to provide information and evidence for problem management processes and liaising with Azure support teams, if required. For this, organisations need to implement robust analysis and troubleshooting tools.
In short, if your Azure servers or services go down you need to know what will be affected, for how long, and what impact that will have on your organisation in order to determine and take necessary steps. Among other things, this means knowing what it takes to keep a high availability application operational if X or Y shuts down. And things do go down from time to time.
“Azure has planned, routine outages for maintenance purposes, when servers going offline temporarily. Before this happens, organisations need to know how many more stand-in machines are required to maintain each service in the event of an outage, compared to traditional, on-premise IT,” says Roger Sinel, Operations Manager at Datacom.
Broadly speaking, the skills required for such mapping, monitoring and management include knowing how applications work, how infrastructure works and how they work together. Operations engineers need to co-operate with developers and application specialists to ensure applications run smoothly in Azure through the correct use of resiliency and performance techniques, and by testing and monitoring correctly. What is supported in Azure and what isn’t need to be understood – especially in a hybrid environment. As reliance on Azure increases, along with complexity, automating parts of processes as much as possible using scripts becomes increasingly important.
The recommendations above are among many others made in our free white paper, How to make the most of Azure, which is available to download. It’s based on Datacom’s many years of experience working in partnership with Microsoft on cloud projects of all sizes for a wide range of organisations. These include what is still the world’s largest production SAP migration to Azure, for Zespri International, one of the most successful horticultural marketing companies globally.