We recently hosted a webinar with Microsoft’s Brandon Middleton to talk about the benefits of the Azure platform. Today, we’ll be sharing part two of our conversation with Brandon and Cloud Academy’s Ben Lambert, where we’ll be talking about big data, PaaS, single sign-on, the maturity of the Azure platform, and more.
How does Azure support big data?
Brandon M.: We’ve got a number of things. We’ve got Azure Machine Learning, which basically allows you to train your cluster of computers to learn on the fly. We’ve got really deep integrations with HortonWorks and Claoudera. They both have offerings on top of the Azure marketplace, which is similar to the AWS marketplace, where you can buy apps and different VNs, and spin them up pre-built and pre-configured to fit the environment.
We also have more white glove sales support for folks who outlay large amounts of capital into the cloud. You can get an official, dedicated Azure support person to walk you through deploying that. [bctt tweet=”We’ve got about as much support as AWS does for their big data” username=”CloudAcademy”] and it’s all based around what the customer is trying to do.
Big data is such a broad term that you have to work together to make sure that you know exactly what the customer is looking to achieve and then work together in partnership with them toward those goals.
What are the pros and cons of taking your big data and moving it to Azure?
Brandon M.: Any company has its subject matter experts. But IT, and specifically being able to spin up infrastructure and give that to the groups that need it, is not a specialty or a core competency for most companies.
When you look at how many folks you want to hire to manage infrastructure, a lot of the conversation is around, “Do I want to be in the business of procuring, maintaining, keeping up, patching, etc. for infrastructure, and for software and technology when that’s not even my core business?”
A lot of CFOs are telling their CIOs, “He, if you can get our footprint a lot lighter in terms of the number of physical servers, and the amount of data center cages that we’re in, you could save the company real dollars, and we could focus on what we do best as a company. We can outsource the rest.” That’s where the conversation usually starts.
I’ve seen IT get transformed in some companies to an internal accelerator, where, in partnership with engineering, they create applications that will go on the Azure marketplace, or in the iOS app store, and they’ll try to become the first in their industry to create a tool or something that becomes a new revenue stream for the company. IT isn’t just a cost center, it’s also something that can generate revenue and generate some returns. It’s the same when it comes to big data.
What about the challenges? Do you hear people say, “We’d love to get onboard with Azure. We’d love to move all of our stuff to the cloud, but …” What is that ‘but’ that they follow it up with?
Brandon M.: It could be a number of things. Usually, a company is not going to rip out anything. They’re going to try to integrate and migrate so that it coexists both on-prem and in the cloud for some amount of time. To be able to do that in the enterprise, you need a clean Active Directory environment (users and groups). You need to know what kind of access policies you want in place, who gets permission to access what resources. We need to replicate, basically, that structure from on-prem into the cloud, and run that in a hybrid fashion.
Sometimes, they don’t have the staff to do it or the expertise, so a lot of my job is to interface with the Microsoft ecosystem partners who actually have the capability to translate the customer requirement, to configuration and deployment, and maintenance, and manage service of that job.
Sometimes the “but” is that they don’t have the people or the skills to do it, and other times they want to test it out first before going all in.
We have another team that is great at doing proof of concepts, and even most of the partners in the ecosystem have week- to month-long engagements where you can show the customer and the entire IT team and staff the delta is between what they’re doing today and what they could be doing in the future. Highlight the simplifications, highlight the economies of scale, and just how your team of IT professionals would be able to scale themselves out so much better and so much more than going to a hybrid or more cloud environment.
What PaaS offerings are available, and what is the overall maturity level on the Azure platform?
Brandon M.: There are lots of different ways to think about platform as a service (PaaS). When I talked about Active Directory, moving it from on-premise to Azure, one of the cool things that we did with that platform is to extend it so that it pre-integrates with applications from single sign-on and multi-factor authentication perspectives.
Watch this short video, taken from Implementing Multi-Factor Authentication on Azure Course, to learn what multi-factor authentication means.
To make the Azure Identity Platform platform as a service, we integrated over 2,700 different SaaS apps so that you can basically sign in once, and then do single sign-on, and then multi-factor authentication. Either use your thumbprint on your iPhone, or use an iris scanner, or biometric strip on your laptop to be able to do away with our fickle brains, and forgetting usernames and passwords across all of the different applications we use at work. That’s one example of using the Active Directory Identity Platform and putting it into the cloud and offering functionality as a service to end customers.
Another platform as a service thing that we’re doing is around [bctt tweet=”being able to write once and push out across multiple different mobile platforms” username=”CloudAcademy”]. We’ve integrated Xamarin into Azure, which allows you to connect your visual studio on our team foundation server to an Azure instance, and basically, have all of your developers pushing up and checking in and checking out code. Then once that’s done in C Sharp, for example, you can translate that into mobile application updates across Windows Phone, across iOS, across Android. So we’re taking the video platform, tied in with Azure, and the functionality of being able to push out code more quickly because we’ve only written it one time, not three times or four times across multiple mobile application platforms.
Those are a couple of examples of platform as a service and things that we’re doing with the big concept of an Azure Cloud, but making it specific to identity, making it specific to mobile.
In terms of security, what does Azure offer that’s similar to a VPC?
Brandon M.: When we talk about security, think of VPCs as more like a redundancy to make sure that you don’t lose data. For anything that is put into the cloud, we back up three times and then replicate that across to another region so there are literally three copies locally, and then another three copies in another region. This way, in the event of a disaster in one region, your data is still replicated and stored somewhere else.
There are a lot of different guidelines around how you should do high availability, how you should write applications. This isn’t even in the cloud, this is just how you should do application development and building from a security perspective. It’s just how to do good quality application development, and it just happens to be on top of the Azure platform.
That’s how we look at it. We want to make sure that all of our customers’ data is safe, which is why we have the compliance that I talked about earlier. In terms of application architecture and in case you forgot to control save something, we have Autosave, as well. So, if you’re writing in Visual Studio and you leave a session open, and you forgot to press save, no worries. Autosave has you covered.
How do you feel about single sign-on making us more vulnerable to security breaches?
Brandon M.: That’s why multi-factor authentication is so important. The stance of Microsoft is to assume that all passwords, no matter how long they are are able to be compromised. Instead, it’s a lot harder to steal someone’s thumbprint. We’re not there yet, and that’s why we’re assuming that anybody can do username and password, and moving on to that second thing is the safest that you can get for right now.
If you had just one set of credentials and someone hacked it, that would be horrible. However, it’s not the case that someone could also find a way to take a stolen set of credentials and then pair it with something that is personally identifiable to you, i.e. your face or your finger, or some code that can be texted to your own cell phone. That second factor, and in the future, maybe even the third factor, will be really, really important to pair with the username and the password.
The cloud has helped us solve things like high availability. What are some of the other things that the cloud is going to help us solve that we aren’t already thinking about?
Brandon M.: This is something that all of the big three are thinking about right now, and if you look at Cortana on Windows 10 devices, or Alexa on the Amazon side, having our machines learn about what’s going on around them and about who we are, and then being able to provide insight will be life-changing, I think. Think about how advertising is done right now. The best that they can do is put stuff up to the right of your searches, to make assumptions about how likely you will be to click through to that thing and then buy it.
If you’re using your compute power as a human being to do life, I think, as opposed to using a computer just to do your job, then we [bctt tweet=”the have ability to have personalized experiences” username=”CloudAcademy”] that computers and technology will be able to offer in the very near future. That’s why our machine learning algorithms are super important, that’s why intelligence is such a hot topic.
When you think about the internet of things and the sensors that different companies are putting out, a lot of it is for learning the environment. Once they’ve completed the data-gathering stage and the have a big enough set, I think they’ll be able to start to derive some insight and say, “Hey, you know, it looks like you left your phone at home. We noticed that your car is starting and you’re leaving without it. You probably won’t want to drive away,” etc. There will be lots of different, crazy things that computers will be able to provide like, “You’re double-booked for your 12 p.m. meeting” and remind us that we need to cancel one and inform the other parties. These are very simple things that will be super helpful.
Check out part one of our interview with Brandon where we discuss the advantages of migrating to the public cloud, Azure support for Linux, comparing Azure with AWS and Google Cloud, HIPPA compliance, and more.
Ready to get started with Azure?
Some great Azure Learning Paths to start from