11 Feb Cloud Infrastructure Integration
We bring you a series of short blogs, quickly readable and always on the essence of things. Feel any need to go more in-depth? Just contact us to schedule a Glueing coffee talk.
Nowadays most companies already use a multitude of cloud solutions, either standalone or already integrated with other cloud or on-premise applications. And whenever cloud is a structural part of the IT roadmap, far-fetching decisions have to be made on the cloud infrastructure, being the important fundament for a long term cloud-enabled IT vision.
Cloud delivery models
The three well-known cloud delivery models have become widely established:
- Infrastructure as a Service (IaaS) providing raw computing, storage, networking capabilities and a pre-installed Operating System of choice as part of the set-up;
- Platform as a Service (PaaS): offers on top of an infrastructure, pre-installed and maintained features like databases, web-services, integration stacks and related development environments (IDE’s). When using PaaS you don’t need to manually configure scalability for example, you just need to write functional code to be executed by the PaaS platform.
- Software as a Service (SaaS): offers a full suite of functionality in a specific business domain. (Line of Business – LoB) like CRM, ERP or more niche feature offerings. Theoretically the only thing you have to worry about is configuring out-of-the-box functionality most often limited to things like workflows, reporting etc -so assembling LoB features into business workflows.
Companies that exploit cloud to the full normally use a multitude of the above service delivery models with a well-thought over integration strategy. Based on business priorities it’s continuously balancing between the usage of IaaS (maximum flexibility, lowest vendor lock-in, biggest DevOps overhead), SaaS (best time-to-market, no customisation, no OPS) and the somehow middle ground of PaaS.
Cloud deployment models
Next to cloud delivery models there is also a decision to be made on the cloud deployment model. Will it be public clouds, private clouds, with- or without on-premise applications or a mix of all? Architectural design and decision making becomes very important because based on these choices the core infrastructural and connectivity fabric is created – the structure that holds everything together.
With cloud solutions being widely available in many flavours, architecting and designing for cloud integration is one of the main but often underestimated challenges. From an integration perspective several interesting market developments are worth mentioning.
- An important development is the upcoming Connectivity-as-a-Service offerings of larger Telco and core Internet providers. More and more data centre connectivity between cloud offerings can be ordered on the spot and paid based on actual usage, made possible by Software Defined Networking (SDN) and Network Function Virtualisation (NFV) and as such, offering more stable and secure connectivity than public Internet. Of course, similar functionality can be achieved with self-managed private network (VPN) topologies and mutual-authentication solutions for IoT and application to application purposes.
- A promising direction is the availability of so-called IaaS/PaaS packages for on-premise installation like for example Azure on Premise also called Azure Stack. This opens up the possibility to transform on-premise infrastructure into an on-premise IaaS/PaaS Cloud with all the standardisation and operational benefits.
- Related to the previous subject the developments of the open source cloud application platform called Cloud Foundry shows a far-reaching cloud-agnostic development. This initiative basically abstracts the Cloud as infrastructure and decouples applications from the underlying Cloud IaaS/PaaS. Such a cloud abstraction layer supports not only a multi-cloud IT strategy but also run-time multi-cloud deployment tactics. Imagine that daily cloud computing resources are scheduled across multi-vendor clouds based on availability and pricing aspects (day-trading of computing resources or insurance companies that shop for additional computing power for monthly resource-heavy prolongation processes).
- The mentioned Cloud delivery models (IaaS, PaaS, SaaS) start to intertwine by new offerings of the larger Line of Business Cloud solution vendors. A company like salesforce.com already offers since long a Platform as a Service, force.com and also SAP nowadays offers SAP Cloud Platform. With these environments, customers are able to develop and run new innovative applications on a trusted and secure platform, close to the core solutions including all necessary connectivity, scalability and security features. They generally come with well-defined service directory (API’s) shielding the core SaaS offering, app-store facilities for application sharing and low-code IDE’s for agile development.
- And last but not least and very important from an integration perspective are the so-called dedicated PaaS environments for integration (iPaaS) like SAP CPI, Mulesoft and Apigee. These platforms provide out-of-box functionality for messaging, queuing, api management, data mapping and transformations, routing, orchestration and most often come with a full set of standard integration between major cloud solutions. Being the integration hub those iPaaS solutions also become a key element from operational perspective with full monitoring, scaling and throttling capabilities.
With this brief overview on cloud infrastructure integration we emphasised on the importance of well-thought over cloud fundament. In our next blog we will share some Cloud Data Integration insights.