re:Invent 2018 Reactions
I've been looking on a good topic to post about before I can get into more example posts and with re:Invent going on this week I thought it might be worthwhile talking about a few of the announcements I'm excited about.
AWS Transfer for SFTP
A lot of the data that I handle at work is given to us by vendors/partners as well as clients. While I would love to have them be able to drop the data into S3 a lot of tools and processes revolve around older protocols like SFTP. With AWS offering a managed SFTP server this will be a good bridge between the old world and the new one. I particularly like that the service integrates with external authentication like LDAP and Active Directory. This will allow not only other parties to handle provisioning the accounts but password reset build around those authentication resources.
EC2 Instances (A1) Powered by Arm-Based AWS Graviton Processors
Back when I first started college I began my academic career studying Electrical Engineering. Ever since then I've enjoyed tinkering with electronics whenever I can. Part of the joy of tinkering has meant a fondness for ARM processors. To me, they always seemed so approachable as someone who's dreamt of building their own computer for so many years. When I heard that AWS had announced the a1 instance family I was very excited. Not only because of that fondness for them but for more practical reasons as well. Things like ARM docker images being built by GitLab, integration testing on Android-based instances, and others make me excited for the future of this and other ARM instance families.
As a side note, I was able to spin up a Docker Swarm cluster on a1 instanced within 24 hours of their being announced. I'm hoping to have a blog post about that ready shortly that also discusses some other ideas and a few I've mentioned here more in depth.
Firecracker – Lightweight Virtualization for Serverless Computing
This particular announcement I'm most excited about because I'm excited to see what the community with do with it. I can envision local versions of Lambda being developed as well as the project being integrated into other projects or services.
AWS Ground Station
Anyone who follows me on twitter (you do follow me, don't you?) knows that I'm a big fan of everything New Space/Space 2.0/what ever you want to call it. If anyone is looking for someone to pay to build and launch a CubeSat feel free to drop me a line. Anyway, this announcement to me means that a big player in the cloud space is ready to take New Space seriously. I've seen several people wonder why this service would be offered. To me, it doesn't seem all that surprising. Jeff Bezos is personally funding a rocket company. In addition, I know of one other company that is leveraging AWS and had modeled their process in a similar way. My only hope is that AWS takes “cloud” computing literally and we see a new region announced: leo-1
Amazon FSx for Windows File Server
Like I mentioned in the section about AWS Transfer for SFTP a lot of the workloads I see on a day to day basis are based off older technologies. I've also seen many based on Windows technologies (I'm more of a Linux fan/user). Having a Windows-based file share means being able to more easily interface with those processes that aren't able to adapt to new AWS-centric technologies like S3.
Control Tower
As someone who's spent two and a half years working on the AWS platform, I think I can say this with a level of confidence: Best practices are hard. Setting everything up, monitoring, all of the various tasks. The fact that AWS is giving an out of the box way to accomplish this hopefully will lead to more and more of these best practices being adopted and all the benefits associated with them reaped.
Amazon Textract
Recently here at work, there was a discussion about the possibility of using AWS’ Machine Learning technologies to extract information from forms that would be sent to us by customers. We ended up going in a different direction but the idea of being able to scan documents and extract the results got me thinking. I personally would use it to take care of receipts and things like that. With Textract all of that is an S3 bucket, a Lambda function, and a scanner away. Receipts, forms, bills, a plethora of items could be scanned, organized, and tagged.
AWS Outposts
With all these posts about AWS, it might surprise people to know that I'm also a sucker for physical servers too. There's just something about having a big, loud, power hungry stack of computers that lets you know work is getting done. I've always secretly wanted to run my own personal cloud based either out of my home or a local data center when the power bill gets too high. Projects like Openstack have meant that that was a real possibility. With the announcement of Outposts, this dream of running a cloud locally is even more possible! In reality, I can see this being used in a lot of places. Manufacturing, Education, the list goes on. My only concern is how much is it going to cost for me to have us-joshs-cloud be a reality.
AWS Managed Blockchain
I've had an issue in GitLab for the blog for a post about developing a CI/CD pipeline for Ethereum contracts. One of the biggest headaches I've been trying to wrap my head around has been how to deploy to the production network and public test networks. Testing is easier thanks to the folks who developed Ganache but the public networks are a different story. I've debated docker containers and AMIs and how to keep the startup time for them down to a minimum and other complications. However, with AWS Managed Blockchain I expect a lot of those headaches to go away. With this service, I'm hoping that AWS removes a lot of the operational headaches of running an Ethereum node and allow for more rapid development on the platform.
AWS Lake Formation
Earlier this year we undertook an exploratory dive into migrating our process from its current state to that of a modern data architecture. This involved the setup of a data lake and exploring various ETL tools. While our particular use case we were exploring didn't quite fit I can say that the concept definitely stuck with me. What also stuck with me was the complexity of such an undertaking. Based on what I saw with Lake Formation I'm eager to see what headway AWS has made into simplifying the process of moving to a data lake strategy.
This is turning out to be the longest post I've written for any of the iterations of this blog so I'm going to leave it there for now. I'll write the rest of my reactions in part two to be released shortly.