Cloud CDN should have the option to give back a generic URL instead of needing a Load Balancer that requires us to configure an IP
Right now Cloud CDN requires the configuration of a Load Balancer, which costs $18 a month. Furthermore, I am given a global IP that I then need to setup on my own subdomain.
What I want is just a way to put a CDN in front of my bucket. I don't mind using a generic url name such as obfuscatedbucketname.googleapis.com/path/to/my/object5 votes
The customer wants to know the usage and charging for every machine type not individual vCPU and memories.
Since October 1st 2018, most machine types are billed as individual vCPU and memory SKUs. The n1-standard, n1-highmem, and n1-highcpu machine types are billed by the individual vCPU and memory usage in two separate SKUs rather than billing by each machine type.
We start to receive a lot of complaints from our customers. Would you please change back to bill by machine types?4 votes
Google Cloud VPN should also support client server VPN for end users to connect from their win\mac clients.
Meantime provide steps to build a basic Windows VPN server \open VPN etc, with LDAP support.5 votes
At present the IP range for BGP peering in a partner interconnect is 169.254.x.x set by Google Cloud. This seems that cannot be customized.These has some problems:
- I need to do monitoring from the peering device to the VMs in the VPC. This monitoring cannot use the 169.254.x.x address range as source IP that the sockets will use by default.
- traceroute not working.
- maybe this list is not exhaustive.
AWS and Azure do support specify a custom IP range for BGP peering.1 vote
Block cursor is hard to work with. Especially when using magnifying on browsers (for example 125%). Please give us option of I-Beam type cursor if it is not hard to implement.3 votes
Make generated firewall rules read-only to prevent editing them accidentally5 votes
A user build its project by adding multiple VPC, subnets, compute resource, network services and other VMs across various regions. A dashboard which provides an intuitive pictorial representation of the complete project would be very helpful. It could be similar to representations you use across your documentation pages. For eg:
The topology may also include below items:
Ip addresses configured on each resources
Vpc peering,VPN tunnels, interconnect etc.
For larger projects, you may provide both minimal and expandable views.
Each vpc/subnet/region/resource on the topology should be a clickable item which would display further attributes associated with it.1 vote
Add a Summarization Service - that could produce summarizes in all different types:
Add a BDI/Multiagent Service - that could be used for gametheory/argumentation, fault tolerance as part of natural computation, reinforcement learning steps. Such services could be useful for advertising platforms, detecting faults, security, chatbots, etc.1 vote
Add a Complex Networks and Social Analysis Service - so its is possible to visualize huge network graphs, support for partition queries, and to combine that with geometric deep learning, bayesian analysis, random forests, and markov models (MCMC)1 vote
Add a QA service - that allows for question/answer permutations and retrieval on search and query as thought vectors on data1 vote
Add a NLG service - this service should allow for someone to type a natural language query that goes beyond static sql but allows for inference at the query plan and get back a context-specific result either in form of a table output or a report. Point here is to avoid having to use sql to query the database but instead to use english sentence as a means of query.1 vote
Add a Performant Transcoder and Compression Service - that could could be used to transcode videos, images, graphics, audio in streaming mode with existing gcp services. Additionally, provide a compression service that could allow for conversion between various file and data formats - anything2triple, avro2json, rdf2jsonld, avro2jsonld, cvs2json, doc2pdf, jpeg2png etc1 vote
Add a Commonsense Reasoning Service - that could connect to google knowledge graph, graph store, and support event calculus for analogical reasoning over data1 vote
Add a Distributional Semantics Service - to support word embeddings, their derivatives and the like, which could be used for topic modelling, seo optimizations and the like....e.g topic2vec, sense2vec, concept2vec, phrase2vec, doc2vec, word2vec, glove, lda2vec, etc1 vote
Add a Search Service - similar to AWS Elasticsearch and Cloudsearch that can have a logging/indexing with cloud storage and connect to graph store3 votes
Add a Graph Database - Tinkerpop and W3C/SPARQL compliant as an alternative to: Stardog, Janusgraph, Neptune. Must support sharding and use of cloud storage for backups.62 votes
One of the fears everyone has is that we'll start learning on GCP and we'll get a bill for $700,000 at home when a bad line of code spawns up 2million nodes on earth and mars.
I'd like to buy $500 increments in Google Cloud Credits. Then when i hit that limit, you cut me off. This allows to learn safely.86 votes
Interesting idea… we’ll share it with the billing team.
Currently it is not possible to connect a Cloud VPN gateway to an on-premises VPN gateway behind a NAT (https://cloud.google.com/vpn/docs/resources/troubleshooting#gatewaysbehindnat).
This is particularly problematic as changing the network topology or changing the HW cannot simply be done on-premises.
I suggest to support this particular setup to avoid to create a VM with strongswan or similar.
It's easier to setup, maintain and configure.
CloudCDN to allow external origin outside of GCP. Currently the only options are GCS or Managed Instance Group4 votes
- Don't see your idea?