Cloud services: the integration problem
The development of cloud services has taken place so rapidly that is nearly impossible to keep abreast. The first successful applications delivered from the cloud not only enjoy rapid development, but the number of new tasks, capabilities and applications that are now being adapted for the model continues to amaze and delight. How about chatbot-as-a-service? Or AI-as-as-service?
According to Janakiram MSV of Forbes, the first wave of cloud computing is attributable to the big platforms, featuring the likes of Google App Engine, Engine Yard, Heroku, and Azure delivered Platform as a Service (PaaS) to developers. This was followed, by Infrastructure as a Service (IaaS) where users would provision virtual machines and storage themselves.
The third wave of cloud was centred around data, he reports, from relational databases to Big Data to graph databases, cloud providers offered data platform services covering a wide range of offerings. Whether it is AWS or Azure or GCP, compute, storage and databases are the cash cows of the public cloud.
“The next wave that would drive the growth of public cloud is artificial intelligence (AI). Cloud providers are gearing up to offer a comprehensive stack that delivers AI as a Service,” he writes.
However, the problem with the increasing sophistication of existing cloud services, and the rapid development of news, is the old problem of integration. How can organisations cope with the lure of the new, delivered with the convenience of the cloud, and the old in the form of legacy systems, or application stacks that need modernisation, but for the moment, still work.
All of which is taking place in a feverishly growing market. A whip around the principle analysts and data trackers finds the global cloud computing market is forecast grow to more than $200 billion (€180.5 billion) in 2019, with cloud initiatives expected to account for the vast majority (70%) of all tech spend by 2020. Most organisations (80%) organisations are predicted to migrate to the cloud by 2025.
Last year, public cloud adoption grew to 92%. Despite this, according to RightScale, AWS adoption fell by 1%, while all of the other major providers, from Microsoft, to IBM, Oracle and Google, saw increases in market share. Forrester analyst, Dave Bartoletti, has declared 2019 the year of global enterprise adoption of cloud, and with his colleague, Lauren Nelson, asserts that cloud computing will be the best way to turn potentially rowdy ambitions into functional software.
This all attests to the huge momentum behind cloud services, for adoption and expansion. And even as organisations feel that pressure, anxieties remain. The Rightscale State of the Cloud Report 2018 finds that the significant cloud challenges are security 29%, lack of resources/expertise 27%, Governance/control 25%, and managing multiple clouds 22%.
However, the challenges do not appear to be denting ambitions for adoption as among large organisations, 81% globally already use a multicloud environment, though coherent strategies remain rare.
“Through 2020, integration work will account for 50% of the time and cost of building a digital platform,” said Massimo Pezzini, research vice president and fellow, Gartner. “Moreover, the complex challenges posed by digital business transformation require a radical change in the integration technology platform and in the way organisations deal with integration.”
So how do organisations go about tackling the issue of cloud service integration?
Gartner recommends a hybrid integration platform.
According to the Smart with Gartner author Rob van der Meulen, added complexity requires a platform approach.
Imagine, says van der Meulen, an organisation implements AI capabilities to answer customer queries quicker. It is likely, he posits, to require seamless connectivity with customer data. To complicate matters, a new customer-facing IoT initiative means that there is an exponential increase in data volume and velocity stemming from thousands of endpoints, the processing of which, he argues, would require migration to a cloud storage and compute platform. This has in turn, enabled real-time analysis which should also feed back into the AI system for greater business insights and optimal business outcomes.
Traditional integration tools, which tend to be task specific, are unable to deal with this level of complexity.
A hybrid integration platform (HIP) is the “home for all functionalities that ensure the smooth integration of multiple digital transformation initiatives in an organisation,” says van der Meulen.
The introduction of an HIP, he argues, means that organisational models need to change to keep pace, as the traditional IT-controlled and centralised integration team with its “integration factory” model, will need to pivot toward an approach that supports HIP-enabled, self-service integration by lines of business, subsidiaries, application development teams and eventually business users.
Gartner expects that by 2022, nearly two thirds (65%) of large organisations will have implemented an HIP to power digital transformation efforts.
The analyst specifies that a true HIP should span and support four key aspects:
• Personas (constituents): Integration specialists, ad hoc integrators, citizen integrators and digital integrators
• Integration domains: Application, data, B2B and process
• Endpoints: On-premises devices, the cloud, mobile devices and IoT devices
• Deployment models: Cloud (potentially across multiple environments), on-premises, hybrid (cloud and on-premises) and embedded in IoT devices
“A sound HIP strategy begins with an analysis of the integration capabilities already in place in your organisation – in terms of the extent to which they support the four dimensions above,” says Pezzini.
“The next step is to define your organisation’s expected integration needs across three time frames: the next 12 months, the next three years, and the next five years.”
While Gartner is fairly unequivocal about the platform approach, there are others.
According to Geoff Bryant, RVP of Services, APAC, MuleSoft, the key to overcoming integration challenges lies in finding a way to modernise on-premises applications and data sources in a way that enables them to co-exist with cloud infrastructure and SaaS applications.
This is achieved, says Bryant, using application programming interfaces (APIs) to build an integration layer that decouples on-premises data and applications from the systems on which they reside.
“This layer also acts as an intermediary between on-premises and cloud services, allowing applications to remain platform neutral and maintain access to data as they move to the cloud,” says Bryant.
Standardising IT access in this way, he argues, also helps organisations overcome the security challenges of moving to the cloud by enabling them to embed controls directly into the APIs, following a security-by-design approach. A further advantage of eliminating tight coupling between services is that applications can be more easily migrated across clouds and SaaS solutions can be more easily swapped in and out. This reduces the risk of vendor lock-in and accelerates innovation by enabling faster deployment in multi-cloud environments.
As organisations build out this new integration layer, says Bryant, they will naturally form what he terms an application network, which provides IT with an architecture of reusable building blocks that can be utilised to rapidly connect new cloud services. With an application network, organisations can future-proof their businesses and create a decoupled architectures that streamline migration between on-premises and cloud environments in a way that minimises disruption to end users, he says.
According to Bryant, enterprises are subsequently free to embrace new opportunities as they emerge, rather than being tied to decisions made in the past – a crucial capability given the dynamic pace of change in the SaaS market. By creating an API layer on top of its on-premises systems, firms can unlock and expose data and application functionality across multi-cloud environments, paving the way for rapid innovation.
Ultimately, the fluidity that businesses achieve with an API-led approach is key to ensuring multi-cloud strategies are successful in delivering the desired efficiency, agility and accelerated innovation benefits, Bryant asserts. As the cloud market matures, it is becoming apparent that cloud migration will never be ‘complete’. There will always be a need to evolve and adapt. The application network is a crucial foundation for a multi-cloud strategy, unshackling organisations from the confines of a static point in time and giving them the freedom to explore best-of-breed services in the journey ahead.
A development of this approach has coalesced as the concept of integration-platform-as-a-service (iPaaS). Gartner describes this as a suite of cloud services enabling development, execution and governance of integration flows connecting any combination of on premises and cloud-based processes, services, applications and data within individual or across multiple organisations.
The iPaaS approach is championed by the likes of Dell Boomi and MuleSoft. Writing for the digital transformation news source Which-50, editor-in-chief and publisher, Andrew Birmingham, says iPaaS “helps businesses and organisations achieve integration across cloud-to-cloud, cloud-to-on-premises, on-premises-to-on-premises and business-to-business environments.”
The implications for cloud service users and consumers are clear: the expected complexity of environments, coupled with the increasing sophistication of the cloud services themselves require tools and methodologies beyond current integration capabilities.
Whether adopting an API-based approach, the iPaaS route, or Gartner’s full on HIP, organisations need a coherent approach and an appropriate toolbox to ensure that the chief cloud integration anxieties of security, governance, compliance and control are met and appropriately addressed.
What is certain is that appetites for cloud services will not be reduced any time soon, and as digital transformation ambitions become more concrete, yesterdays measures will not yield the success required for tomorrow’s services.
However, the perennial challenge will remain, namely security. Securing such complex environments will continue to be a challenge, as skills remain in short supply, and nature of the job changes constantly.
Ensuring data does not leak, or inadvertently go where it shouldn’t, are still concerns, as cloud misconfiguration and privileged access attacks figure highly in breach reports. Measures such as biometrically secured log-ins, remote security management, and encryption of hardware are likely to be brought to bear to ensure that any cloud service, and the data it processes, can be adequately protected.
Identity and connectivity issues
Integration between on-premises workloads and cloud services is reasonably straight forward provided two things are looked after:
- Identity – Lack of identity management across the enterprise not only creates significant barriers to productivity but also encourages the development of a patchwork quilt of identity and role mappings which equals security risk. Simple actions like federation of Active Directory domains into the Cloud and the use of Identity and Access Management (IAM) services such as SailPoint can help greatly in complex environments.
- Connectivity – It goes without saying that as cloud services are remote from the on-premises environments connectivity is a must but I wonder if people who are new to services like Microsoft’s Express Route realise that this is not just a generic big data pipe but has to be used with particular workload design to get the best throughput from it?
In Tekenable, we have tested and experienced what works and what does not over Express Route. Careful planning is required, particularly for hybrid solutions where database elements are on-premises with significant cloud-based demand.
Multi-cloud implementations must by necessity, if inter-cloud portability is a requirement, avoid using the very services that typically differentiate cloud providers and deliver that last 10% of competitive advantage to the cloud adopters.
For example, TEKenable has used Microsoft Cortana Intelligence Services (AI) on Azure to deliver amazing customer service experiences in a hybrid cloud solution. That would not have been possible, or at least would have been much more complex, if we had to allow for portability to AWS or Google Cloud for example. Using virtual servers and container services does not save you from vendor lock-in if you use differentiating cloud services, but as long as cloud vendors remain competitive most companies will take that last 10% competitive advantage.
Peter Rose is CTO of Tekenable, tekenable.ie
The cloud journey
We are observing a growing trend – the focus on the journey to the cloud. Many enterprises are not considering the end destination, be it hyper scale providers or private cloud environments, as the measure of success rather the journey to get there.
The improvements and efficiencies gained by going on the ‘cloud journey’ delivers real value in itself. By planning the journey carefully and by ensuring you have the right management and orchestration platforms for your environment you can actually ensure a hybrid environment. This encompasses existing infrastructures – be it on premise or in a private cloud, alongside some functions such as DR and back up moving to the public cloud. Also new line of business applications originating from the cloud (cloud native) work well to ultimately deliver against the SLA you are most likely under from your business.
If you have clearly documented what you currently operate and have a view on how this can be optimised plus understand how you can take advantage of Cloud services for the growing requirements from your business you can start to build out a management layer or cloud management platform (CMP) to deliver against pre-defined outcomes. As with any IT environment a strong management capability is usually at the core of success.
What we in Triangle, are increasingly seeing is a hybrid approach to cloud. This will come as no surprise to anyone, as there has been a hybrid approach to IT for a long time.
However, what it presents is one of two things:
• An increasingly agile way to run your IT operations and deliver new services to your business with the benefit of insight packed data to leverage for sustained improvement
• A complex environment with infrastructure and services at different levels making your operation challenging to manage and difficult to budget for. It can also promote shadow IT as more and more cloud services target the end business user. Great for speed of execution but it needs to be governed
Paul Flavin is commercial director of Triangle Computer Services
To deconstruct and modernise
As cloud adoption evolves, more and more organisations are looking to realise the well documented benefits that this provides, such as reduced cost, scalability, business continuity, speed to market, collaboration efficiency and ease of integration.
Furthermore, as the need to break down data silos, enable transparency, provide ubiquitous access to products and services, and optimise business processes increases, cloud requirements have progressed from being primarily infrastructure driven to the provision of fully integrated Software as a Service (SaaS) solutions.
So how do companies get the best out of their cloud strategies in this inherently complex hybrid world, where on-prem’ and fully hosted solutions need to coexist and interact with each other seamlessly, whilst also adhering to increasing levels of governance and regulation around data and security?
Organisations can define standardised application design principles for new integrations, based on Domain Driven Design with a bounded context outcome, providing loosely coupled services where each component is a full but contained application that is focused on producing a single business outcome or result.
They can look to embed common, standards-based specifications like REST and JSON for both new and existing integrations to enable a common integration layer between different components and solutions.
They can also look to systematically deconstruct and modernise relevant legacy and monolithic application stacks, and rehouse these in a container-based integration platform using cloud native standards. This will enable applications to continue to fulfil existing functions while removing dependencies on legacy, tightly coupled components allowing them to be integrated across environments and exposed to new applications and integrations.
Adopting these principles can provide the bridge between systems embedded in organisations legacy stacks and newer cloud-based capability, which empowers organisations to integrate using an agile, scalable solutions that promote reuse and automation utilising more progressive continuous integration and continuous delivery methodologies.
Paul Mahon is senior technology consultant with Singlepoint
Integration game changer
Cloud integration is a game changer. It is a system of tools or technologies that addresses the age-old problem of business silos: applications and data that were historically separated into, for example, CRM, finance, or logistics systems that typically did not talk to each other. Before, companies had to download the data they held into Excel and configure it to get meaningful insights. Now, cloud integration is a way to connect those systems and use the data they have in a whole new way.
Any business looking to move to the cloud must have an integration strategy from the very start. Most businesses do not use the cloud as a single, homogenous whole, but as a hybrid of on-premises, hosted and cloud-based systems. Cloud services like Office 365 or Hubspot, are just as at risk of being a data silo, unconnected to any other system in the business. Unless the business considers each cloud service as part of a wider integration plan and takes steps to unify the data in an orchestrated way, it is hard to gain efficiencies or insights.
By integrating the systems, the business reduces complexity, removes unnecessary double entry of data and moves away from scenarios where different departments own different systems. When linking its data together, the business should also consider what information it wants to have within easy reach. What is the one pane of glass view, either of the business, or of a particular customer, that will help improve decision making?
Fortunately, cloud integration projects are relatively quick to implement between four and eight weeks. There is a rapid return on investment but also what we call a return of interest. That is the value to the organisation by having their people focused on the right tasks, having a standardised quality of data, and deriving insights from that data to help them to make better business decisions, faster.
Oliver Surdival is head of Cloud Services with Arkphire