Local presence, global business is becoming a norm today. Advances in cloud technology and the availability of new infrastructure services, have reduced the time for a business to launch in new markets.
Madhusudan Shekhar, Principal Technical Evangelist, Amazon Internet Services Private Limited outlines the strategy to take customers directly to the AWS platform and deliver cloud services. He speaks about new services like Amazon CloudFront, AWS Direct Connect, Amazon Route 53, and Amazon SageMaker that are making it easier for AWS customers like Freshworks to provide services to their customers globally – from India. Global customers like NetFlix could quickly launch their services in India over the AWS platform.
DC: What are the segments you are targeting? How strong is your presence in India and Asia?
Madhusudan: We launched our Asia Pacific (Mumbai) Region in the year 2016, and we have grown to hundreds and thousands of customers in India. Our business spreads across enterprise, small and medium businesses, and of course, startups — even young developers. So these are the key audiences we go after. And from our public sector business, we go after education which is universities like Gujarat Technological University, IIITs, and the quasi-government organisations that operate in India. So we reach out to our customers across this entire spectrum in this space.
We have 18 regions globally. In India, we have one region in Mumbai. From the 55 Availability Zones globally, there are two in India in which customers can actually initialize infrastructure and operate our cloud based services.
In addition to that, we have Edge locations, which are in essence responsible for accelerating. At the end of the day when you provision or engage with infrastructure resources on the cloud, one of the things we want to be able to manage… and especially in the media space, or when customers are transferring content like uploading files, downloading files. They (our media customers) have the ability to be as close as possible to the customer. We support that through what we call our Edge locations. We have 114 Edge locations globally, out of which 5 of them are in India. Amazon CloudFront is a server which accelerates the content delivery, but Edge locations also provide a number of mechanisms for, let’s say, enterprise customers to land their content near their enterprise networks in AWS. So expanding on that, we have a service called AWS Direct Connect.
We have customers like Freshworks who have built the software and servers in India. They are selling that to America, Australia and New Zealand. So they actually go top of infrastructure on our region in Sydney, and they are now able to expand the services.
DC: Can you tell us more about AWS Direct Connect?
Madhusudan: AWS Direct Connect enables our customers to bypass the internet and go directly to AWS. If you are a large enterprise and you have large data centre in India, you can use Direct Connect to go from your data centre into AWS infrastructure directly, which is what we call as VPC environment. You will not go through the internet pipe but through your own MPLS connections.
AWS Direct Connect actually does the handshake for you. Direct Connect is sold by our partners. We have five such locations in India, so for example an enterprise customer based in Bangalore can land the network in Bangalore, the Edge location is in Bangalore. And then subsequently after that, they can provision across the entire 18 regions and all 55 availability zones in any infrastructure that they want, and all of that will travel over Amazon’s internal network.
DC: It is a trend these days for content delivery networks to have their own fiber networks running across the globe. What is AWS doing in this space?
Madhusudan: We have our own fiber (network) running throughout the world, and it is one of the largest fiber networks. It is bi-directional, so you could say if there is a fiber cut at one end of the transfer route, it automatically reroutes and gets connected to the other side. On top of that, we have services like Amazon Route 53, which is available for you to route traffic to the nearest points of business.
Amazon fiber network runs globally between our regions and our availability zones. This kind of infrastructure has enabled customers in India to take the servers that are built in India and offer it to the world. For example, we have customers like Freshworks who have built the software and servers in India. They are selling that to America, Australia and New Zealand. So they actually go top of infrastructure on our region in Sydney, and they are now able to expand the services. They don’t have to do a massive build up, as it is a sales organisation. They can start offering servers to Australian customers. They get their sales organisation going in Europe, close to the customers.
And then there are global companies like Netflix, who offer services throughout the world, when they entered India, they could use our platform to actually offer the same services. That’s the sort of the capability and it fundamentally enables whether you are a large business or a small business. The ability for you to take your services back and offer it to the rest of the world.
We are able to penetrate new markets very quickly because of our infrastructure and capability across these markets. Next comes the scope of new products and capabilities we have been launching.
DC: How are you able to penetrate new markets so quickly and with such momentum?
Madhusudan: We are able to penetrate new markets very quickly because of our infrastructure and capability across these markets. Next comes the scope of new products and capabilities we have been launching. Post Amazon re:Invent (the annual AWS event for developers and partners, where new products are announced), you must have seen the amount of acceleration we have done on machine learning and AI. We have been doing AI for the last 20 years, and we have taken that and packaged it. This is not very different, if you look at how we evolved IT services and IT capabilities as part of cloud.
Amazon in 2005 learnt a better way of packaging technology and offering it to its customers because we did it for ourselves. We took that repackaging and offered it as cloud and we got a six-year head start with that. And now we have been practicing machine learning for 20 years for our business and our own efficiency. We have taken that construct and packaged it as either machine learning services or packaged it as platform services that customers can actually build. And you can build and recommend the systems the same way we built it; you can build image recognition systems the same way we built it. You can build machine learning platforms that go on IoT devices, on drones, on any kind of factory automation devices etc and build the same thing that we do in your platform. That’s something we have offered to customers today. The acceleration on that is high and we launched with ten different algorithms. Today we are at 13 algorithms with Amazon SageMaker, which enables customers to rapidly innovate on top of that. We had added a substantial number of capabilities on our vision, voice today — Polly as a service, has 24 languages.
Editor: Amazon Polly is a text-to-speech service that turns text into lifelike speech, allowing businesses to create applications that talk, and build entirely new categories of speech-enabled products. Amazon Polly uses advanced deep learning technologies to synthesize speech that sounds like a human voice.
It offers dozens of lifelike voices across a variety of languages, you can select the ideal voice and build speech-enabled applications for different countries.
We have been practicing machine learning for 20 years for our business and our own efficiency. We have taken that construct and packaged it as either machine learning services or packaged it as platform services that customers can actually build.
DC: Can you give us one example of an Indian customer who is using Amazon Polly?
Madhusudan Shekhar – Policybazaar creates highly customised call back for instance. The customer goes on their site and tries to buy an insurance policy. When one buys an insurance policy one needs to submit a bunch of documents, such as KYC, income proof, three months’ bank statement etc. Now one of the key challenges is, if a customer uploads two months of bank statements and misses one month. The classical call centre scenario is the agent has to digest all this in their head and they actually need to have a conversation with the customer — you cannot automate that without the agent that easily. Because all of these machines have to be customised. You have to say: Dear Mr. Customer, you have applied for this policy from Bharti AXA. And for this health insurance policy you need to submit three months of income proof. You have uploaded two months and you have not uploaded the third month. This is the kind of message.
The string can be constructed out of the data systems that they have.
Now with Amazon Polly they are able to convert that to a full voice file and actually have that land on the customer without an agent. The software actually figures out the whole (English) script.
ADDITIONAL INFORMATION
- AWS Asia Pacific (Mumbai) Region was launched in June 2016 and became the sixth AWS Region in Asia Pacific (APAC)
- AWS investments in India include 5 PoPs (2 each in Mumbai and Chennai and 1 in Delhi) which are a part of the AWS global network.
- In November 2017, AWS Direct Connect landed sites in five cities – Denver, Phoenix, Madrid, Helsinki and Chennai
- The AWS Cloud infrastructure is built around Regions and Availability Zones (AZs), providing customers with an easier and more effective way to design and operate applications and databases, making them more highly available, fault tolerant, and scalable than traditional single data center infrastructures or multi-data center infrastructures