Dynamics of Private-Public-Multi Cloud Installations

When mapping their IT strategy, companies need to evaluate opportunities and challenges presented by cloud technology and many businesses find themselves with a mixed private and public cloud setup. This whitepaper explores important issues with multi cloud scenarios, with a focus on the monitoring of multi-cloud solutions.

Why Companies arrive at Multi Cloud Scenarios 

The uptake of cloud technology by large, high-profile global companies, its popularity within IT departments, its promise of ‘almost free’ (or at least low cost) and the rise of cloud-based solutions have predictably forced companies to scrutinise and reconsider their traditional legacy systems and setups. Every significant company must develop a cloud strategy defining if and how they should move their existing setup to the cloud. They will likely start the transformation process implementing a private cloud solution on-premises infrastructure to achieve a scalable and more manageable system. However, as business activities expand, it becomes inevitable that the need for more computing resources or external software services grows to the point where the company faces a dilemma – expand the private cloud resources or acquire resources and services from a public cloud provider.

There are many benefits of using a public cloud, like the flexibility in deploying new environments within minutes or quick reaction to peak workloads, to name just a couple. This results in better employee experience and efficiency. Public cloud providers offer solutions that enable quick business transformation journeys and faster go-to-market requirements. Public cloud providers are generating and offering more and more new cloud based and cloud native services that reinforce technological innovations and modernization. Public cloud providers have also already addressed many regulatory and legal requirements and continue to do so. Driven by these very compelling arguments many companies will chose to expand their private cloud infrastructure with public cloud offerings. This results in an ever-increasing number of multi-cloud setups. 

Multi Cloud combines the best of both worlds 

Private Cloud - 

• Security aspects, regulatory and legal requirements - Certain systems and data registers may not be placed into a public cloud environment. Possible reasons are constraints regarding geographical placement and the required physical access (e.g. emergency infrastructure, co-location requirements, bandwidth, data restrictions).
• Financial aspects - For example, the hardware that is already owned and still not depreciated (and software if it cannot be reused) can be used for running a private cloud design.
• Services or functionalities - Some services cannot be provided by a public cloud provider or are uneconomical to adapt for a public cloud scenario.

Public Cloud - 

• Scalability - Complete flexibility to react to resource needs instantly without requiring oversized hardware for everyday business.
• Dynamic creation of environments - Development and test activities benefit from unrestricted and automated creation and destruction of environments on demand.
• Innovative services - Bringing in and leveraging innovative, external cloud-based services is easy.
• Evaluation and proof of concept (POC) - Evaluation of products in cloud setups is very easy and can usually be done without financial commitments outlasting the duration of a POC.

These examples show why a multi-cloud solution is often used, as it is necessary for companies for compliance reasons and they can also get the best of both worlds. However, even though a multi-cloud solution can bring considerable benefits for a company, it also brings inherent risks, specifically with monitoring.

Current and Target State of Multi-Cloud Monitoring 

Importance of Multi Cloud monitoring for a hybrid situation

An ideal and solid multi-cloud monitoring system needs to have attributes of high availability and centralisation. A quick win might be to choose a Monitoring-as-a-Service (MaaS) solution where a predefined, pre-configured and scalable monitoring service is hosted in a public cloud. Solutions from all major monitoring providers can be found as instantly deployable services within public cloud providers’ marketplace sections. Another solution consists of creating and managing your own centralized monitoring system installation where either your local infrastructure or private/public cloud hosted infrastructure stack supports performance monitoring of infrastructure and application layers. Additional attributes include a fully customisable visualisation interface and flawless compatibility with mobile platforms as a must. Finally, a wide support of API and webhook connectivity options towards mostly used and popular ITSM tools is an important attribute to facilitate an effective alert/event management.

Implementing a multi cloud monitoring solution 

The general approach of implementing a new multi-cloud monitoring solution should contain the following steps and milestones -

• Identify the pain points and missing information that would protect the business.
• Together with relevant business units determine what needs to be monitored. Outline how they benefit.
• Select a suitable monitoring solution based on requirements.
• Set up an implementation strategy - MaaS or own installation.
• List metrics, key performance indicators (KPIs) to monitor and events to log.
• Develop an incremental implementation strategy based on criticality and value.
• Deploy the solution for a trial run and setup alerts for high priority events and metrics. Establish and test connections to external tools used as part of the ITSM and initiate business reporting.
• Complete the setup of metrics monitoring and event logging.
• Control the metrics for accuracy and tune thresholds to the desired targets.
• Decommission legacy monitoring systems.

 How to choose the right multi-cloud solution

There are many solutions on the market, and it is difficult to make the right choice. During the selection process, you should consider the following questions -

• What type of monitoring are we seeking? The most common types are network and security, infrastructure availability and capacity, application performance and availability, web performance and user interaction monitoring, business performance metrics (e.g., the number of processed documents, transactions, sales results, etc.).
• What is the preferred management technology and required information granularity?

– Agent monitoring – typically designed for a specific platform and vendor, can collect more detailed information.

– Agentless monitoring – uses standardized protocols and log analysis. In general, ready-to-deploy solutions are hosted in private/public clouds and usually have great built-in agentless monitoring possibilities.

• How many devices and applications are to be monitored?
• What is the company’s licensing strategy? Would it rather invest, rent, or use an open-source solution, potentially contributing own development to implement specific functionalities?

Many companies are running workloads, applications, and systems in a distributed landscape with a combination of public cloud providers, private clouds and legacy on premises infrastructure. Without the right solution, the monitoring of all these environments can become a useless overhead in the best case and a danger to the business in the worst. Therefore, we regard a multi-cloud monitoring solution as an important infrastructure item in this scenario. Implementing multi-cloud monitoring will require investment, but it is a crucial step towards operating a stable business and achieving lower total costs.

By Florian Nemling, Senior Consultant, Martin Rehker and Alan Benson, Managing Principal at Capco

Read more IT news here

Read products news here