September 19th, 2018 | by Krzysztof Krzemień
Cloud Developer Days 2018: A New Place to Be
Table of contents
When we think about the cloud, we often contemplate daydreaming. A state of relaxation and doing nothing, when thoughts come and go. The Cloud Developer Days 2018, a brand new conference created by the veterans behind .NET Developer Days that took place on 28th-29th May, is a place where daydreaming turns into knowledge and inspiration. Keywords – lower maintenance costs and higher security for solutions in the cloud.
The CDD conference focuses on solutions and innovations in the field of cloud computing. During the two days, there were 30 sessions and keynotes. With speakers like Peter Hedges from IBM, Radu Vunvulea from Microsoft, Bartosz Chojnicki from Mindbox, or Paweł Rzepa from SecuRing, one would expect a mountain of value; it was delivered. That amount of knowledge deserves a book, not a humble blog post, so we have decided to focus on two exciting topics – automation of infrastructure in a cloud and artificial intelligence (AI), and machine learning. What will come out of these solutions and what does all of this mean for the ways entrepreneurs do business?
Saving Time and Resources
Cloud computing is trending among entrepreneurs and IT professionals. Not all remember to prepare a good setup stage, which is crucial for seamless maintenance. The infrastructure can be so complicated that they are unable to remember all of the setup options. They can also lose track of the virtual machines variety. The business is scaling, therefore you need more servers and more advanced configurations. Your favorite administration panel is not only inefficient and can’t handle the business’s needs and react to changes, but it also lacks quality and can’t maintain the recurrence (standardization) of the infrastructure itself. Automation is the cure, but how to implement it for the solution to be easy to use and scalable?
Naturally, solutions are coming directly from vendors, such as CloudFormation Template from Amazon or Azure Resource Manager from Microsoft. There is nothing wrong with these technologies, but there can be a time when you decide to reach for more diversified infrastructure, coming from different providers and that can generate challenges. A real-life scenario for using multiple vendors can look as follows.
Read also: Cloud Developer Days 2019: An Automation Tool for Azure
Your application’s production version runs quicker on Amazon Web Services (AWS) but your internal servers are cheaper when maintained on Azure and their efficiency is not that important. A single solution to rule them all could be beneficial (no, we are not using a magic ring). It can help automate the creation of the infrastructure while supporting multiple providers at the same time. This will radically simplify configuration and lower the cost of having few solutions in day-to-day use.
The name of the solution that could handle this and other, similar scenarios, is called Terraform and it comes from HashiCorp. This tool allows for the creation of homogenous infrastructure configuration and it comes as a code. It also allows for versioning and storage in the version control system. Terraform also enables the creation, modification, and erasure of infrastructure, no matter what particular service from whichever vendor is in place. It can proudly show off itself by having multiple modules compatible with all widely known cloud computing vendors. We had the opportunity to see this beast during the CDD conference and it looks very promising; especially if you want to save time and reduce the cost of resources within the cloud.
Less Hassle with A.I. and Machine Learning
AI and machine learning are other hot topics in the IT community. The CDD conference served as a showroom for platforms working in the cloud, utilizing both technologies. In the age of free-of-charge technologies, there’s a vital question about the validity of using the cloud’s computing power and infrastructure with tools for the solution creation process. Let’s imagine a case here:
A company makes integrated circuits for subcontractors. For quality assurance this company has planned on using image processing technology, able to spot anomalies during the manufacturing phase. The realization of this assumption can be fulfilled in two ways:
- The company will build its physical infrastructure along with the software capable of image processing, based on free technologies.
- The company will build a part of physical infrastructure, that only takes samples for analysis. The computing and the software part will happen with the help of cloud-based services.
The first model means, that the company in question generates costs by manufacturing and implementation of the quality assurance ecosystem. Let’s not forget the responsibility for stability, reliability, and maintenance of this big solution. The final aspect involves the time needed for manufacturing and implementation.
The second pattern requires the entrepreneur the resources and time needed for manufacturing the infrastructure for sample analysis. Algorithm training for anomaly detection with the computing infrastructure is moved to the cloud. The benefits are obvious – the reduction of responsibility and the hassle involved with keeping the full solution are a win.
One of the main solutions that can be helpful is Microsoft Azure. The A. I and Machine Learning section has the following services:
Batch A.I. – the platform for deep learning and A.I for models. Batch A.I. supports cooperation among Azure’s virtual machines. It includes the newest GPUs from NVIDIA. Alongside with the flexible coding model, it supports easy implementation and scaling of computing resources, effectively supporting any requirements tied to models.
Cognitive Services – a platform for recognition services, delivering algorithms for speech, image, and text processing. It also supports data mapping.
Bot Services – a platform supporting the creation and implementation of intelligent bots. At the conference, we had the chance to see Microsoft Bot Framework with SDK Bot Builder.
Machine Learning Studio – a tool for building, testing, and implementing solutions from the field of predictive analysis, delivered in the form of a drag-n-drop designer.
Microsoft Azure delivers a complete set of solutions allowing for a quick and effective design and exposition of created services alongside delivery of the exact amount of computing power for data processing. The pathway looks like this:
- You can create a model using the Machine Learning Studio. • You put a data set with processing configuration by using modules available inside the platform • You choose a learning model • You choose a grading and evaluation model • A training data set is enough for the model to learn
- Then you can configure the computing infrastructure available in the cloud. • You choose several processes, GPU, memory, and the number of machines responsible for calculations.
- The final step is the exposition of the model, by creating a www service.
Read also: Cloud Developer Days 2019: Kubernetes and Istio
The Cloud Developer Days Future
With the very first edition of the Cloud Developer Days behind us, it’s too early to say how this very informative and welcome event is going to blend into the landscape. What we can say for sure is that we can’t wait for the 2019 edition, announced at the end of this year’s event. We went, we saw, and we left mightily impressed. The team behind .NET Developer Days did manage to bring a talented and knowledgeable set of people, with intriguing ideas to contemplate and practical solutions to implement.
Spot-on, spot-on.
