Suggestions
Michael Krotscheck
Identity Expert, PaaS Architect, Keep it Simple.
Michael Krotscheck is an experienced professional in the tech industry, currently working as an R&D Software Engineer 6 at Broadcom. His LinkedIn username is indeed krotscheck.1
Professional Background
Michael has over a decade of experience in designing, developing, and delivering cloud-based solutions.1 He is known for his expertise in identity management, PaaS (Platform as a Service) architecture, and his "Keep it Simple" approach to problem-solving.57
Career Transitions
Prior to joining Broadcom, Michael worked at VMware. His time at VMware came to an end due to the company's acquisition by Broadcom.4 After a brief job search, he secured a new position, presumably at Broadcom, as indicated by his current role.5
Areas of Interest
Michael shows a keen interest in various aspects of technology and industry trends:
- Cloud Security: He has attended events like CloudNativeSecurityCon to stay updated on industry developments.6
- Artificial Intelligence: He has pondered the potential impact of generative AI on C-Suite roles.2
- Healthcare and Compensation: Michael has shared thoughts on the importance of healthcare benefits as part of total compensation packages.3
Professional Outlook
Michael appears to be an engaged professional who actively shares insights and participates in industry discussions through his LinkedIn profile. He values networking and staying current with technological advancements in his field.
Highlights
I've been noodling on a problem recently, and was wondering who here has experience with hot-swapping a large, read-only database table.
Assume, for a moment, you've got a single table, with a single column of roughly 600M unique 4B records, which is replaced wholesale a few times a day. There's no incremental updates, and the query pattern only ever checks individual records - no listing.
What databases exist that can pre-build the index and table "elsewhere", can hot-replace the table from that pre-calculated index, and - most importantly - are open source? What approaches have you taken?
My current thoughts on this:
- We could just run a sidecar with the DB already wrapped into it and link it on a new port. No fuss, no muss, but works better in k8s style deployments instead of individual instances.
- Would SQLite work? It's read-only, after all, and could be treated just as a volume mount.
Introducing OpenREST! 🌟
One of my favorite bodies of work at VMware was the creation of an opinionated RESTful API standard, a document built over months in close collaboration with my colleagues in the SecureState team. We argued, we revised, we voted, we revisited and considered use cases, and finally built a strong consensus that relegated API details to the same closet that "tabs-vs-spaces" got shoved into.
Since my layoff, I've had the time to publish it openly, and after a few weeks of heads down writing (and desperate scouring of my memory, because I didn't have access to the source material anymore), I'm happy to publish it to the world!
It's high time that we stop arguing about the "right" way to implement an API layer, and get on with focusing on our business needs. Share, and share alike!
https://lnkd.in/gtwqDV8T
#API #REST #OpenREST #Collaboration
