Brands
Discover
Events
Newsletter
More

Follow Us

twitterfacebookinstagramyoutube
Youtstory

Brands

Resources

Stories

General

In-Depth

Announcement

Reports

News

Funding

Startup Sectors

Women in tech

Sportstech

Agritech

E-Commerce

Education

Lifestyle

Entertainment

Art & Culture

Travel & Leisure

Curtain Raiser

Wine and Food

YSTV

Amazon Web Services

Amazon Web Services

View Brand Publisher

At AWS Cloud Day 2019, AWS experts discuss how their Cloud solutions can help organisations scale as happy customers agree

At AWS Cloud Day 2019, AWS experts discuss how their Cloud solutions can help organisations scale as happy customers agree

Thursday August 08, 2019 , 13 min Read

As part of its mission to put Machine Learning in the hands of every developer – Amazon held AWS Cloud Day 2019 on July 19, 2019 in Bengaluru to discuss topics like AI/ML, SAP, IoT, Migration and Cloud Architecture, and how Amazon’s Cloud solutions are helping startups scale. The event saw AWS solution architects, industry veterans, CXOs, CTOs, IT managers, facilities heads and startups coming together to share customer stories, and explore how they can use AWS technology to innovate and build solutions for their businesses.


How to build a pattern of resiliency


Madhusudan Shekar, Head - Digital Innovation, AWS took to the stage to talk about how to build a reliable system. "In the digital world, one of the most important things you need to build for is resilience. Being down and unavailable is not acceptable. You lose a customer's trust,” he said.


Irrespective of what system you build, there will be failure, but you need to approach failures with systems predesigned to handle them. Resiliency is fundamentally built by ensuring that you build your design around high availability. "It is the ability for a system to handle and eventually recover from unexpected, extremely rare conditions," he said.


Madhusudan said that there are four layers to take into consideration while building a high resiliency system - infrastructure, network and data, application and people. "People are very important. You need to have a go-to person for any kind of failure, but also need to build resiliency in lieu of them, in case they are not available.


He also spoke about the concept of Blast Radius, or the impact when something fails. You need to analyse if the failure is something you can absorb, and what customers, functionalities and locations would be impacted. Eventually, your application should be able to support the impact of the failure. "Firefighters practice with fire. The only way to understand fire is to walk into it. When the actual fire hits, it’s like any other day for them. Manage failure every single day, so that you can go about fighting it when it hits."


He further spoke about Availability Zone (AZ) in AWS and how AWS has multiple data centres with servers that are interconnected with 3 AZs, which enables fault-tolerant applications. He also said that there are three common issues with databases - scaling, replication and backups. Startups can leverage AWS regional services such as Amazon S3, Amazon DynamoDB, Amazon Aurora, Amazon ELBs, which are purpose-built for the cloud and designed to withstand these issues.


According to him, once you build a good system, customers come your way and your business starts to scale. "You will scale when you’re successful, trustable, have good sales and marketing."


Anatomy of AWS services


The session covered the core services of AWS and how the company has grown from a few thousand customers when it began, to over 2.5 million active customers on its platform today. In 2008, they had 24 service updates, and it’s 20 times more in 2018. In the first five years of launch, they were present in just five regions. Today, they have 65 availability zones among 21 regions, with 187 points of presence. They have recently expanded to the Middle East and African subcontinent as well.


Availability zones are one or more data centres separated by a meaningful separation. AWS has a fully isolated infrastructure with one or more data centres. Most regions have three or more availability zones, and each of them are connected with microfibers.


On the core services of AWS, from around five systems in 2017, AWS has 175 unique systems today. Whatever the use case, their broad set of compute instance types cover it with four characteristics - CPU, memory, storage, network performance.


When you’re architecting the system, it’s important to understand the science behind the nomenclature. They have pricing models for each instance, the default one being on-demand instances, where you pay for the compute capacity by the second with no long-term commitment. If you’re sure of the capacity, you can go with reserved instances, which are steady-state usage with a one-year or three-year commitment. Finally, there are spot instances, which are popularly used by customers as they are flexible and fault-tolerant.


The session emphasised on the philosophy of AWS and what makes them unique. AWS believes in putting ML and AI in the hands of every developer globally. Their ML and AI services are in three categories. The first level is for expert Machine Learning practitioners who are already used to the framework level. There are only a handful of customers in this section. The second level comprises developers and data scientists who want to manage the environment while training and deploying ML models. The third layer is for a class of users who want to integrate AI capabilities into their existing/new applications.


The talk concluded with the seven different aspects that set AWS apart, the topmost being security, among breadth and depth of services, experience, global footprint, ML capabilities in hierarchical structure and lifecycle, ecosystem of database partners, and enterprise leadership, among others.


Customer success stories


AWS has several customer success stories, of organisations who have leveraged AWS Cloud services and scaled their businesses. Manjunath K G, CEO, Kenscio Digital, a global award-winning digital marketing agency, spoke on how they are able to improve email marketing for organisations by partnering with AWS. He said that email is becoming a prominent medium for marketers to target their audience, and it’s the easiest way to communicate with their subscribers. "Be it transactional, communicational, informing password changes, promotional, emails are the way to go."


Some of Kenscio's products are built on AWS, and as a partner, they are helping brands on Amazon to make the best use of Amazon infrastructure. He spoke about the different ways in which email marketing could go haywire, because of a lack of understanding. This includes the kinds of emails to be sent, understanding how email service providers regulate emails, filtering emails, analysing metrics, and so on. "Infrastructure is a continuous challenge when it comes to filtering email and email metrics. If customers don't repeatedly open your email, it goes to spam. We have a lot of tools to filter emails. This has improved inbox rates by 20 percent to 30 percent, open rates by 20 percent to 30 percent, and click rates by 30 percent to 40 percent,” he said.


Kenscio has the right infrastructure to understand data. You get insights for every campaign, you can check repetition when you send emails, update your email threshold policies, and so on. "Your database is built into the platform with dashboards, push notifications, customer journeys etc.," said Manjunath.


Naveen Dachuri, Co-founder & CTO, Yulu Bikes is another customer of AWS. He realised that the average speed of big cities during office hours had reduced below 6 kmph, and this was not due to inefficient use of personal vehicles or cabs. Also, 14 out of 20 most polluted cities in the world belong to India, and one-third of this pollution is caused by vehicles, which can be solved if clean modes of commute are used.


He wanted to solve the problem of traffic congestion, make short distance commute more efficient and greener, and make cities more breathable. "I wanted to make the process more efficient using micro-mobility as a service, and that’s how Yulu was born,” said Naveen. Yulu solves the problem of last-mile short connectivity through urban micro-mobility, by building Yulu Move, Yulu Miracle, and Yulu Magic.


Naveen shared how they're driving sustainable mobility through innovation and a lot of research. "We're building products in the shared mobility space that users love. Our vehicles are light-weighted, easy to move, with less breaking parts. There's no particular target segment for their product, as everyone uses it.”


On how they build new features, Naveen said that they build it on the go. "Charging the vehicle was one of the challenges, so we built a charging infrastructure." Yulu technology is simple, secure, scalable and unified. In terms of infrastructure, they have built smart vehicles, charging stations, and smart power control systems. Their smart and intelligent systems are data-driven, and they use AI and ML for prediction, routing models and intelligent alerts.


On why they chose AWS, Naveen says that the setup is fast, there is reduced hardware cost and procurement cost, and analytics is available. "We do all the testing using AWS products. Everyone in the team is comfortable and gets trained to use this dedicated resource."


ML and AI applications on AWS


The next session highlighted why AWS is perfect for AI and ML businesses. Through a demo of image analysis by Amazon Polly, a text-to-speech service that uses advanced deep learning technologies to synthesise speech that sounds like a human voice, it covered AWS's recommendation systems. AWS’s mission is to distil down information so that anyone can use it. As long as you know how to code, you get the power of recommendation systems and voice analysis in your hands.


AWS has the broadest and deepest set of AI and ML services with 200+ new features since January 2018. They are solving the toughest problems and support all popular frameworks. Their approach is customer first – build a system or application based on what they ask.


The talk further delved into their AI and ML solutions, which are divided into three parts. The first is AI services that solve existing problems like image recognition, analysing transcripts etc, for which no experience is required. The second is ML services for individuals who have done the initial POC with AI services, and still find inferences or data that are not good enough. This can be used by people who are familiar with Python and Amazon SageMaker. The third service is ML frameworks, which can be typically used by data scientists, people who want to figure it out on their own.


Re-inventing SAP, Data Lakes and analytics on AWS


The session that followed covered how people traditionally ran workloads on SAP and how you can reinvent SAP on AWS. AWS has worked closely with SAP since 2008 to certify SAP offerings on AWS. When they release new instances, they work with SAP to certify their solutions on AWS. With the agile platform, what used to take months to build is now done in an hour with AWS.


The talk delved into how you can deploy SAP using Quick Start reference deployment, or automated templates that will spin up products like HANA, Netweaver and S/4 HANA. The benefits are the speed of deployment and predictability because of automation. There are certain blocks for running SAP like networking, storage and file shares, but there are corresponding equivalents on AWS to overcome these blocks. When you're running SAP, resiliency, availability and security is very important for customers, and AWS takes it very seriously. They have 50 certifications and numerous security features.


The session also covered CloudEndure, which lets you install an agent and do block-level replications. You can do SAP migrations using CloudEndure. It has little to no disruptive installation.


APIs are critical for a connected experience. APIs provide an agile, flexible, secure and scalable layer for integrating various business applications and act as an easy-to-consume service for all your user interface and integration needs. The talk concluded with the Amazon Machine Learning stack which has three fundamental stages -setting up the environment and figuring out which algorithms to use; training your model on data; and inferencing or deploying ML models to production. The entire process can be time-consuming and difficult, but Amazon SageMaker automates the entire process, gives you a framework, and lets you focus on building ML models.


Another session covered the topic of how customers want more value from data now. They want a single data store that is scalable, cost-effective, stored securely in standard formats, and analysed in a variety of ways. Cloud Data Lakes are the future.


Security on AWS


Security is a primary concern at AWS, as they currently host data for 1 million active customers, with new verticals and diverse use cases. AWS works closely with customers to achieve security and compliance through secure platforms and cutting-edge services. Their core is the inbuilt infrastructure with 24/7 monitoring systems that follow the CIA (Confidentiality, Integrity and Availability) structure for customers.


The next session delved into the sophisticated AWS services that customers can use to enhance their security posture and focus on their business. With AWS, you inherit global security and compliance controls, and scale with superior visibility and control. There are the highest standards for privacy and data security. You also get to automate with deeply integrated security services and get access to a large network of security partners and solutions. You can access services and tools, enabling you to build compliant infrastructure on top of AWS.


AWS Infrastructure Security services reduce the service area to manage and increase privacy and control of your overall infrastructure on AWS. These include Amazon Virtual Private Cloud (VPC), AWS Shield, Amazon Inspector, etc. They also have a set of Detective Control services which help you get visibility into issues before they impact the business, improve your security posture and reduce the risk profile of your environment. These are AWS CloudTrail, Amazon GuardDuty, Amazon CloudWatch, among others.


AWS Activate for startups


The event concluded with a talk on why AWS is a preferred partner for most startups. There are three main reasons - geographical reach, breadth and depth of service use case, and the way we've innovated in the last 13 years.


AWS's startup business development team builds partnerships and maintains a strong relationship with members of the ecosystem. They build an environment where these partner organisations and their portfolio companies can get appropriate visibility into AWS strategy, direction and value proposition.


According to a survey by YourStory, 90 percent of India’s popular startups are on AWS, and this was because of the benefits that AWS brought to the table. Through the AWS Activate programme for bootstrapped startups and AWS Portfolio programme for seed or Series A startups, they get access to credits, business support, startup spotlight, one-on-one technology mentoring with AWS solution architects, capital connect, PR outreach, and more.


Every startup needs to build a brand while scaling and attract new tech talent. For that, you need to showcase what you’re doing – AWS does case studies for startups who have grown their customer base. The capital connect allows startups to meet VCs in a speed dating format, and they also get to pitch to potential VCs and senior management of companies.


The Launchpad programme by AWS also helps startups launch their product faster. You get discovered through PR opportunities, get direct access to investors, build your brand and connect with potential customers.