Microservices architecture embraces small, self-contained services built, deployed and scaled independently. However, despite the allure of microservices architecture, the monolith is still relevant.

Enterprises need to find a middle ground between microservices and the monolith. A good strategy is to create a hybrid microservices architecture.

What is a hybrid microservices architecture?

To adopt a microservices architecture, enterprises refactor an application to distribute services individually. A hybrid microservices architecture typically contains monolithic application code alongside a collection of scalable, platform-agnostic business components that suit cloud-native, containerized deployment. This approach combines the features of both monolithic and microservices architectures — some parts of the application are built like microservices, and the remaining parts continue to operate in an expected and well-understood way.

Teams that create a hybrid microservices architecture can reap the benefits of distribution without having the need to refactor an entire monolith, saving a huge investment of time and resources, and preventing potential disruption to the app’s performance and functionality.

How to implement a hybrid microservices architecture

A hybrid microservices architecture means implementing microservices only where they provide a benefit:

• Reusable components that have value across the app portfolio;
• Components that need flexible scalability; and
• Features that undergo frequent updates.

The first step in this journey is to split an existing monolith into logical and business components, and determine what components are common or reusable. Analyze the business benefits of using a combination of the two architectural styles in the same application, and decipher which components the software team can still build and deploy as a monolith.

Because microservices offer scalability as a primary benefit, identify the parts of the application that are easier to scale than others. Study the resource consumption of each module. For instance, if your application is designed for multiple concurrent users, you might want to consider making security a separate microservice to scale it according to expected workloads and how much data there is to protect.

The other important factor to consider is how often you update a service or component. A good reason to use microservices architecture is to enable frequent updates without disturbing the other parts of the application. Facilitating these seamless updates often means designing applications for a microservices architecture. On the other hand, data access components of the application may not require — or even benefit from — frequent upgrades. In this case, you can keep a monolithic codebase while the other parts of the application are converted to distributed services.

Key takeaways

You might have thought about moving an existing monolithic application to microservices. However, you won’t be able to convert a monolith to a microservice quickly or easily. There’s a good chance you don’t have the right technologies and tools, resources and developer expertise already in place. Take the hybrid approach and only use microservices when it is required — not just for the sake of using it.

The post Get started with a hybrid microservices model appeared first on Artificial Intelligence.

Hybrid cloud—environments that blend public and private cloud—are a central part of the modern IT landscape, particularly as organizations of all sizes work to incorporate legacy systems into future-looking digital strategies and while meeting regulatory requirements. Although well-established, hybrid cloud is often poorly understood by IT departments, CIOs and partners, and its long-term trajectory is unclear to many that work with the cloud. As Microsoft positions Azure to play in a hybrid cloud world, MSDW reached out to experts for their thoughts on where hybrid cloud is headed next.

Assessing the benefits of hybrid

When it comes to hybrid cloud, many organizations adopt a hybrid approach to keep data under their own control, while gaining the benefits of scalability in the cloud. Adam Mansfield, commercial advisory practice leader at UpperEdge explained:

Using a hybrid cloud, enterprises have the ability to put sensitive data in a highly secure and controlled private cloud environment while having the ability to also place other non-sensitive data in an often more scalable, reliable and cost-effective public cloud environment, like Microsoft Azure, AWS, or Google Cloud Platform (GCP).

According to Dr. Taras Filatov, CEO and founder of blockchain-provider Dappros, the proportion of cloud components relative to on-prem continues to grow with time. To improve hybrid cloud performance, Filatov recommends using Kubernetes—which is still at a low level of adoption—to containerize applications. Additionally, he suggests taking advantage of machine learning and blockchain add-ons as well as leveraging public cloud compliance tools for regulations like GDPR.

In many cases, hybrid scenarios are a way for companies to boost agility while continuing to get value out of on-prem data centers. But in other situations, hybrid is a way to handle poor connectivity with remote locations like rural factories, farms, or oil and gas sites. What’s more, hybrid gives a way for companies to spend time unpacking years of what Andrew Bandera, DMI practice lead and solutions architect, calls “enterprise accretion,” tidying up massive customized systems as a staging area before a move to the public cloud.

The post Where hybrid cloud is headed next, and the role of Microsoft Azure appeared first on Artificial Intelligence.

Medical imaging is among the most popular application of AI and machine learning, and with good reason. Computer vision algorithms are naturally adept at spotting anomalies experts sometimes miss, in the process reducing wait times and lightening clinical workloads. Perhaps that’s why although the percentage of health care organizations that have adopted AI remains relatively low (22%) globally, the majority of practitioners (77%) believe the technology is important to the medical imaging field as a whole.

Unsurprisingly, data scientists have devoted outsize time and attention to developing AI imaging models for use in health care systems, a few of which Google scientists detail in a paper accepted to this week’s NeurIPS conference in Vancouver. In “Transfusion: Understanding Transfer Learning for Medical Imaging,” coauthors hailing from Google Research (the R&D-focused arm of Google’s business) investigate the role transfer learning plays in developing image classification algorithms.

In transfer learning, a machine learning algorithm is trained in two stages. First, there’s retraining, where the algorithm is generally trained on a benchmark data set representing a diversity of categories. Next comes fine-tuning, where it is further trained on the specific target task of interest. The pretraining step helps the model to learn general features that can be reused on the target task, boosting its accuracy.

According to the team, transfer learning isn’t quite the end-all, be-all of AI training techniques. In a performance evaluation that compared a range of model architectures trained to diagnose diabetic retinopathy and five different diseases from chest x-rays, a portion of which were pretrained on an open source image data set (ImageNet), they report that transfer learning didn’t “significantly” affect performance on medical imaging tasks. Moreover, a family of simple, lightweight models performed at a level comparable to the standard architectures.

In a second test, the team studied the degree to which transfer learning affected the kinds of features and representations learned by the AI models. They analyzed and compared the hidden representations (i.e., representations of data learned in the model’s latent portions) in the different models trained to solve medical imaging tasks, computing similarity scores for some of the representations between models trained from scratch and those pretrained on ImageNet. The team concludes that for large models, representations learned from scratch tended to be much more similar to each other than those learned from transfer learning, while there was greater overlap between representation similarity scores in the case of smaller models.

To rectify these and other issues, the team proposes a hybrid approach to transfer learning, where instead of reusing the full model architecture, only a portion of is resused and the rest is redesigned to better suit the target task. They say that it confers most of the benefits of transfer learning while further enabling flexible model design. “Transfer learning is a central technique for many domain,” wrote Google Research scientists Maithra Raghu and Chiyuan Zhang in a blog post. “Many interesting open questions remain, [and we] look forward to tackling these questions in future work.”

The work comes shortly after Google detailed an AI capable of classifying chest X-rays with human-level accuracy. In another recent study, teams from the tech giant claimed to have developed a machine learning model that detects 26 skin conditions as accurately as dermatologists and a lung cancer detection AI that outperformed six human radiologists.

The post Google proposes hybrid approach to AI transfer learning for medical imaging appeared first on Artificial Intelligence.