A Moment Of Reflection…

Jeremy Bartosiewicz, Cloud Architect at Cloudreach, takes a moment to reflect on his time at AWS re:Invent 2018 and the announcements that excited him the most.

As I slowly pack my bags before the new Machine Learning Beta exam at the end of re:Invent, I find myself reflecting on the event itself. With the myriad of Vegas lights, casinos, activities, meetings, and, more importantly, announcements, I always find that this is the best moment to take a breath and digest the veritable feast of information that has built up over the past five days.

So (deep breath), what got me excited about re:Invent this year?


Many of our customers came to visit our booth and play our game Two Blocks Away. It was a great opportunity to have maintained discussions around its implementation and share ideas about how the new AWS releases might impact them. It’s always a privilege to be seen as a strategic advisor, and not just a body shop (looking at you, GSIs).

AWS Transit Gateway

As someone practiced in the art of networking, I think the new Transit Gateway is very exciting. I have implemented similar designs in many solutions with appliances. I like to ruffle some feathers with my thoughts, so I will be the first to evaluate patterns and implementations whilst leveraging this service alongside integrations with other cloud providers. I know many would appreciate this, so stay tuned!

Custom Run Times & ALB support for AWS Lambda

As someone who has focused much on serverless technologies recently, I am as excited as the rest of the community about custom runtimes and ALB support for Lambda. Of course, its only speculative (but highly likely) that already supported run times will be highly optimised but it’s always a pleasure to see AWS listen to what we all want (thank you Amazon!). The ALB integration with Lambda will make life much easier when chopping and changing sites and services.  

Hybrid Cloud solution

The final thought in my mind sits around the Outposts service. I think many have been expecting something like this, and it has caused a cascade of interest and mixed emotions at the conference. Some see it as Amazon admitting defeat in the requirement for hybrid Cloud. Others see it as reaching parity with other cloud providers.

My view is slightly different. AWS has sweat blood and pulled teeth trying to avoid complex hardware which can be used for hybrid cloud. It feels this is because they have focused on their Cloud-based offerings; making sure their services are well thought out, performant, and implemented such that this offering is only relevant to very niche use cases (microsecond networking requirements and seriously stringent compliance and data regulatory needs).

Their timing with this service is spot on. Most naysayers have managed to make workloads work in the cloud, and this seems like a great solution for the final few, which have been filtered very carefully by delaying the release of such a capability. Again, only once the hardware and capability is really fit for purpose (being product focused) would it make sense to release into the wild.

Again, thanks to AWS for keeping us on the straight and narrow!