Not necessarily. Combined with the move to compressed storage, really they're shifting the cost from storage to compute. Storage heavy customers may see their cost go down. Probably this just better aligns the pricing with the cost model.<p>You might also see a bit of savings by using Autoscaling which wasn't available before.<p>Net/Net it depends on the specific usage pattern whether your cost will go up or down.<p>The per-GB price for compressed is higher, but more than offset by the fewer GB being measured, depending on your compression ratio. If the per-GB price doubles and your compression ratio is 8:1, that's a 75% reduction on the storage side.
It’s not just BigQuery. Pricing for cloud services is increasing across every provider. Azure increased prices in the EU recently <a href="https://news.microsoft.com/europe/2023/01/05/consistent-global-pricing-for-the-microsoft-cloud/" rel="nofollow">https://news.microsoft.com/europe/2023/01/05/consistent-glob...</a><p>For AWS, one of the people behind <a href="https://ec2instances.info" rel="nofollow">https://ec2instances.info</a> showed that the price/performance ratio for newer instances is rising, <a href="https://github.com/patmyron/cloud/#compute--memory-unit-prices-by-virtual-machine-type">https://github.com/patmyron/cloud/#compute--memory-unit-pric...</a> meaning cloud users are paying more for less.
So they have decided that the growth period for GCP is over and now it needs to squeeze all the profits it can. I thought with all that search money pouring down, they would bet a little longer on GCP growth.<p>I remember many large companies that migrated to GCP explicitly cited BigQuery cost effective performance. This should change that calculation significantly.<p>But if GCP loses all cost effectiveness, why would companies chose it, knowing that unlike AWS, GCP products might be priced like a loss leader for some time and can increase any day.<p>Also, if Google becomes content to be just a small cloud provider, do they just bet their entire future on search. What if search gets disrupted (by something like a next gen LLM).
Cloud providers raising prices is unsettling. Teams invest many person-years into building software on top of cloud services assuming certain prices. There is a lot of lock in there. Although it is hard to predict to what extent they will pass along future cost savings, or if new cheaper options will come available, it has been safe to assume they won't jack up prices on any given workload.<p>By violating that trust, it will make teams wary of building on top of Google's higher-level services where there is the most lock-in, which unfortunately for Google, is also where the highest margins are.
GCP's biggest problem is trust. People (rightfully) don't trust that GCP won't do a rug pull with their production workloads. GCP has done that with both pricing and killing off services.<p>If you are an infrastructure provider, the first thing people are paying you for is to be a stable foundation. If there is no stable foundation there is no trust.<p>Its really hard to imagine GCP fixing this without new leadership. Hire someone from AWS who knows what they are doing. AWS doesn't do this shit. AWS goes so far out of its way to not break customer workloads.
Annual flat rate plan is going from 2.3c per slot hour to 4.8c. On demand pricing increasing by 25%. Seems like they are trying to squeeze the users who are most locked in.
These threads are part of the canon of "stupid things the hackernewses will say about market economics". It's truly amazing. There is no economic system where a company decides to raise prices based on how much money they need to fund some other thing. That is just beyond stupid. It's a multi-supplier market and prices are set via equilibrium price discovery.<p>Please, HN. Try to be less stupid.
When you run out of resources on an elastic cloud: "here's a bill for 2-5x what you thought it would be, good luck paying it!"<p>When you run out of resources on metal: "your application is slowing down, becoming unresponsive, and falling over, good luck getting it back up!"
Current employer spends a lot on BigQuery. This change is a ~100% price increase for us.<p>As a direct result, we are drawing up plans to abandon BigQuery, and then GCP entirely.<p>What an own-goal for GCP. Not even Oracle has treated me this badly.
Bigquery is one of the tools i’ve been most impressed with the last 20 years. It just works so much faster than I assumed it would.<p>Unfortunately for Google I think this will push some folks to self host some of the OSS alternatives, but maybe not.<p>When I think of tools I'd want to pay for outside of barebones VMs in a cloud, the things at the top of my mind are Newrelic and BQ.
The sensible thing to do at this point is to have a hybrid setup and being able to move load between Cloud providers and from/to on-prem. I understand this is not always practical, but it is going to be a big lever to keep costs down. A Databricks instance running on Cloud can be probably 10x of the cost of running Spark clusters on-prem. Does it make sense? Well, it depends, if you are spending 1m in Databricks charges it does, if you are spending 100k it does not. But I don't see much of these discussions in IT, I think part of the problem is that Cloud charges are hard to monitor and budget.
It was rather expensive (but I'd still use it again) last year when we migrated a custom solution to BigQuery. Now it's even more expensive? It was easily 10k/month for a medium startup.
We are not a direct BigQuery competitor, but we're building an open-source data platform based on Kubernetes over at Stackable: <a href="https://www.stackable.tech" rel="nofollow">https://www.stackable.tech</a> | <a href="https://github.com/stackabletech/">https://github.com/stackabletech/</a><p>It can be interesting for users wishing to avoid vendor lock-in, there are trade offs of course.
Pieter Levels's tweet [1] from earlier about two of google's biggest earners - search and maps - degrading over time had me pondering what google's response would be.<p>I guess ramping up prices across other offerings is one option.<p>[1] <a href="https://twitter.com/levelsio/status/1641759968394113024" rel="nofollow">https://twitter.com/levelsio/status/1641759968394113024</a>
Slot hour? It's been a year or two since I worked with bigquery but I enjoyed that you only paid for data read.<p>It was such a simple way to think about it.
Flagged editorialized title.<p>Article is PR spam.<p>Actual content is the price sheet:<p><a href="https://storage.googleapis.com/gweb-cloudblog-publish/images/1_packaging_option.max-1400x1400.jpg" rel="nofollow">https://storage.googleapis.com/gweb-cloudblog-publish/images...</a>
Gone are the days when hosting price was dropping every year. Now it's increasing. Blame it on inflation or market downturn or whatever, but a trend that was here since the beginning of the Internet, has now reversed.
Google also just added new SKUs last year related to multi-region storage buckets, too. I wonder if we'll start to see similar price increases from AWS/Azure. Usually the cost of tech trends down, not up.
Like big oil, it sounds like there is enough "infrastructure" out there that allows them to move the price fixing to third party companies that "recommend" prices for stuff.
My immediate answer is:<p>So? It's not like the customers didn't knew this was a REAL possibility. Given the track-record of Google and companies similar to Google, this was a known quantity even at that time.<p>Now pay up.
it was well publicized that Google was turning the screws on its GCP sales staff...maybe this is a put-up-or-shut-up moment for GCP to bring in more money...assuming the sales drives didn't yield