5 ms·
[infrastructure bias] Kubernetes and TensorFlow should count, and are successful. AlloyDB is IMO most likely to be successful (especially since AWS Aurora alr
by rrdharan 4y ago
[infrastructure bias]
Kubernetes and TensorFlow should count, and are successful.
AlloyDB is IMO most likely to be successful (especially since AWS Aurora already proved the market):
https://cloud.google.com/alloydb https://cloud.google.com/alloydb
Since this question seems to be much more about the consumer side, I think both Google Home and YouTube TV are independently considered successful though I have no doubt many people will chime in to note how much they hate either or both of those things.
- plorkyeran 4y agoKubernetes isn't a product.
- rrdharan 4y agoStrongly disagree. Kubernetes is not a Google product. But it is "a" product, and it was successfully launched as a commoditize-your-complement / cannibalize-yourself kind of thing. GKE meanwhile is a Google product, and is successful, and wouldn't exist if it wasn't for Kubernetes.
- dangus 4y agoWhat about GKE?
- lake_vincent 4y agoIt's interesting how legacy companies are not making exciting consumer/commercial products anymore, but infrastructure and technical projects are booming. Google TensorFlow and DeepMind, Microsoft WSL2, Meta AI, etc. Also worth mentioning the many quiet efforts to get quantum computing off the ground.
- formercoder 4y agoI work at Google, but just want to say that AlloyDB is really cool.
- davewritescode 4y agoSeems like Google's version of AWS Aurora?
- formercoder 4y agoI don’t know the details of Aurora, but AlloyDB separates compute and storage for Postgres. Masters can scale up and down with no downtime, and storage is provisioned in real time. Only pay for what you use. There is also a columnar engine cache, basically BigQuery, for OLAP queries.
- makestuff 4y agoYeah that is similar to Aurora. Basically a query engine on top of S3. Not sure if it is based on postgres or not though. It wouldn't shock me if it was as I know AWS Redshift is a forked version of postgres.
- formercoder 4y agoRedshift is a columnar database like BigQuery, so wouldn’t be based on an OLTP DB like Postgres. More I work for Google disclosures, but I think BQ kicks Redshifts butt.
- makestuff 4y agohttps://stackoverflow.com/questions/34853204/is-aws-redshift-to-postgresql-the-same-as-aws-aurora-to-mysql https://stackoverflow.com/questions/34853204/is-aws-redshift... Looks like I was mistaken. It uses some parts of postgres 8, but it is heavily modified. I agree Redshift isn’t that great though and it is super expensive for what you get. In the past when I have used Redshift I have not been that impressed. Although one nice feature of Redshift is how well it works with other AWS services (COPY command from S3, firehouse can dump data into it, etc.)
- grepLeigh 4y agoMy most basic definition of a product is: Something I can insert money into, which solves a more expensive problem. Kubernetes and TensorFlow are extraordinary technologies - and very important to the work I do daily. I don't think they're products though, certainly not successful products.
- rrdharan 4y ago> Something I can insert money into, which solves a more expensive problem. Right. I would argue that "public cloud infrastructure" is such a product; GCP/AWS/Azure specifically. (See below also) Kubernetes enabled the existence of EKS/AKS/GKE all of which I believe are considered successful. Separately, TensorFlow enabled Azure Machine Learning, TensorFlow on AWS and Google Cloud TensorFlow Enterprise, and generally expanded (IMO) the market for IaaS.
- kro 4y agoFlutter also comes to mind
- wanderingstan 4y agoAs a user, I can say that Google home has become a dumpster fire after starting out rather well. I’ve sworn off buying any future Google hardware.