Enterprise Architecture in the age of “Everything as a Service”

Cloud services such as Infra as a service (IaaS), Platform as a Service (PaaS) and Software as a Service (SaaS) have been widely adopted in enterprises. Simultaneously, a ton of cloud enabled innovation by established players and startups alike, have resulted in a broad array of, off the shelf cloud services. This has ushered in an era of “Everything as a Service” when it comes to enterprise technology. In spite of operating cost and security arguments raised against the cloud services, they provide compelling solution capabilities. Though these services are broadly classified as SaaS, PaaS and IaaS, a closer look at some interesting sub-categories, is necessary to appreciate the proliferation and maturity of these offerings.

Analytics as a Service (AaaS): Pre-built data analytics tools for data analysis, dashboards, and reports. Examples -Tableau Cloud, Looker, Power BI.

Artificial Intelligence as a Service (AIaaS): Ready-to-use AI models and tools, making AI capabilities accessible to businesses of all sizes and technical expertise. Examples - Amazon SageMaker, Azure Machine Learning, Google Vertex AI Services.

Internet of Things as a Service (IoTaaS): IoT solutions over pre-built platforms, connectivity, and data management tools. Examples - Microsoft Azure IoT Central, AWS IoT Core.

Identity as a Service (IDaaS): User authentication, authorization, access control, single sign-on and multi-factor authentication etc. Examples - Okta, Azure Active Directory and Ping Identity.

Data as a Service (DaaS): Data on-demand; Access to pre-processed, curated datasets or customized data feeds. Examples - Dun & Bradstreet Data Cloud, Credit bureaus like Equifax etc.

Business Process as a Service (BPaaS): Business processes such as workflows for accounting or customer service, offered as a service. Examples - Xero for accounting, Salesforce for CRM and Zendesk for customer support.

Communication as a Service (CaaS): Communication solutions like video conferencing, messaging, and email over cloud to enable collaboration without traditional infrastructure. Examples -Zoom, Slack and Microsoft 365 Exchange.

Integration as a Service (IaaS):?Automates integration and data exchange between different applications and systems. Examples - Boomi Integration, MuleSoft Anypoint, Zapier.

There are numerous others such as Security as a Service (SECaaS), Desktop as a Service (DaaS), Network as a Service (NaaS) and Disaster Recovery as a Service (DRaaS), but you get the picture.

The enterprise architecture challenge:

Options in general complicate life. Technology architecture was complex enough before cloud, and now the options made possible by cloud, make it even more complex. There are decisions to be made: Build Vs. Buy; On-premises Vs. Cloud Vs. Hybrid; Multitude of choices for each capability. Working with enterprise architects, solution architects, and of late with many data scientists (especially in context of Gen AI), I see a lot of unnecessary hesitation to use out of the box cloud services and stitching together solutions using these services, as opposed to custom solution development. However, I would like to present few good reasons to give due consideration to cloud-based services in designing enterprise technology solutions.

The cloud services are maturing and evolving fast: The last few years have seen significant maturing of cloud services in various categories. Innovation have been brought about by hyper-scalers who have both scale efficiencies and massive R&D budgets. The innovation and capability maturity has also been driven by some very strong startups who have been enabled by generous venture capital funding and talent. The advances in Big Data, AI, Gen AI, AR, VR, IoT and the likes, have led to even more innovation and specialized cloud services. This trend is only going to gather even more momentum.

Bespoke applications take time and talent: The economics of developing deep technology solutions make bespoke application development less attractive for many organizations. Many of these new capabilities need significant R&D investments. Talent may not be readily available in the market. Development timeframes may be long, and risks involved may be too high for the business to absorb.

Matching the overall value of cloud services may be a challenge: Considering the sophisticated choices and capabilities offered by cloud and competitive pricing, unless there are strong internal competencies or compelling business reasons, the argument for bespoke application development is progressively getting weaker. Even in case where you don’t have a single cloud service that meets your needs out of the box, capabilities exist to orchestrate cloud services and combine them with bespoke applications or other cloud services to meet even sophisticated business needs. A good example is the use of LangChain to orchestrate Gen AI solutions.

This is not to say that cloud services or their combinations are the solution to all technology architecture problems. Let us look at what situations better suited for SaaS (and cloud services) and when Custom Application development is warranted.

Instances when out of the box cloud services are the better choice:

• Standard functionalities that are not core to competitive differentiation of your business.
• When speed to market is a priority.
• When you are new or have relatively low maturity in a business or technology capability.
• When TCO (total cost of ownership consisting of development and operating costs) are favourable. Typically, SaaS is a no brainer for low volume, complex tasks, if fitting solutions are available. Same is?case for growing businesses or those with fluctuating needs.
• When looking to reduce fixed costs or technology footprint of your business.

Situations better suited for bespoke development:

• When your organization has a high level of business or technical capability in an area.
• Solution involves lot of internal IP and competitive differentiators including proprietary data.
• When your business requirements are unique or complex.

Another key point is that technology solutions need not be SaaS Vs Custom Application. Hybrid solutions are also possible. Most SaaS applications offer APIs (Application Programming Interfaces) or Pre-built connectors. Integration Platform as a Service (iPaaS) platforms provide tools and templates for building and managing integrations across various applications and services. You can also write your own wrappers applications that stitch together multiple SaaS and internal workflow applications.

Given this backdrop, it is today’s need that enterprise architecture gives due consideration to out of the box cloud service alternatives both as solutions and solution components, in designing enterprise grade technology solutions. Architecture choices needs to be comprehensively assessed covering capability, fit, risk, upfront and ongoing costs, security, internal expertise, need for competitive differentiation as well as current future business agility.