Cheaper cloud spanner starting price
It would be great if Cloud Spanner had a cheaper starting price. I listened to Deepti Srivastava on https://www.gcppodcast.com/post/episode-62-cloud-spanner-with-deepti-srivastava/ talk about customer feedback on pricing and how they felt that Cloud Spanner provided incredible value.
I 100% agree that it provides incredible value, and the performance that you can get from a single node is great and comparable to a MySQL VM that costs the same amount. However my needs are much more modest than 10,000 queries per second or 2,000 writes per second. I'd be quite happy with a node that provided 1/10th of the performance at even 1/4 or 1/2 the price of a single node.
High availability, reliability, and transactions are all important for my app, I just don't have a high throughput to start with. My choices are: start with a more modestly provisioned Cloud SQL instance and then rewrite my app when I get enough customers to warrant the $600/month for Spanner, or start with Spanner and essentially over-provision it.
Thanks, we’ve heard this feedback from a number of different directions, and the team is aware of the desire to see something like this. We’ll update this request in the future if there are new offerings to announce.
Felipe Lascaray commented
I think Cloud Spanner would have to implement a fully managed version with a pay per use model like Firestore
Kote Isaev commented
I guess that some cheap offering of Spanner can have special limitations meaningful for small companies and startups. For example, the Spanner Mini (lets name this way):
1) max data size is 1 TB. Option to switch to full-blown Spanner if data size reached this limit.
2) 1 or 2 smaller nodes, 1 per availability zone.
3) Smaller nodes mean smaller RAM, 1 or 2 CPU.
This could be an option with price like $50-$75
Nathan Allan commented
As priced, Cloud Spanner is far beyond the reach of hobbyists and early startups. Not only does that exclude a large market segment, more importantly larger companies come from smaller ones, and people tend to not radically change platforms. I advise startup companies, and I would love for them to be able to start on technology that scales with them, but the technology has to scale all the way down to meet them. Starting at nearly free allows for scrappy startups to develop against the platform and get through the stages of nearly zero scaling needs. It's an investment in future profitable customers.
At an absolute minimum, there should be an emulator or local development story so that users aren't paying $600/mo during development.
It would be very good!
Alan Kaifu Lu commented
Before we fully know and use powerful GCP, many investments have lots of risk. So in order to let more companies and developer to deep research and fully use all kinds of function, we need have more attractive policies.
Having more new clients, having more new GCP service deeply integrate into clients cloud systems, having long-term and stable cooperation. All of them are what we most want to see, right?
mark f Edwards commented
i would like to see a "test" version google-spanner. we are trying to connect a framework to spanner and i believe google-spanner should be more economical when being accessed in a development mode.
Daniel Compton commented
Also, it'd be great if there was a separate area in uservoice to give feedback on this.