I want to try two things in google cloud.
Nested virtualization to create bunch of simple office computers with gpu support (nvidia grid).
Premptible render nodes and AI training nodes , to work with some fault tolerant algorithms.
I have activated my account. Requested more GPU all regions increase. Mailed to the support. Tried to chat with technical support.
-- Mail delivery subsystem error. gc-team#google.com
-- Chat did not start, circle swirled forever
What should I do ?
Thank you.
As I have been informed, in some regions, google cloud prefers opening gpu's and some services through their solution partners. I have contacted sales, arranged a meeting. In this meeting, they redirected me to a solution partner. We signed a reseller contract, prices and services are same unless you ask for additional value-added support from local reseller. They gave me a sub-id and registered my company domain and e-mail to gcs. Everything works now.
Bottom line I get from this is:
If you wanna get serious with google cloud you may want to contact sales directly.
Related
I was really satisfied by the billing structure of cloud functions. They only charge during the runtime of a function based on how much resources it consumes. I am looking for a similar solution for my game servers.
I checked various cloud hosting options like VMs but all providers like digital ocean, google, amazon etc. charge even when there is no load on VM (and I completely understand the reason for that).
I am looking for an option where I can deploy my game server and will charge me only when it will consume resources(based on how much resources it consume) and won't charge (or a minimal charge) when the server is idle. It's like it will auto scale based on the current load.
Thankyou in advance for all the answers.
You can try Squids.io(https://squids.io/), a fully managed database service provider allows you to power off when your server is idle, and you can restart it at anytime of your choice. Plus, it charges you no management fees.
I want to start an IoT project with multiple machines (i.e., Raspberry Pis) in different locations (i.e., throughout a city or several cities).
Each machine should be able to send data to a back-end (ex., alive ping, sensor read-out updates, etc.) or receive orders from a back-end and send back response.
I don't want to start from scratch (can't use Amazon or Azure IoT service) - is there a framework that would do this (can register to the service and allow me to check which machines are up/down, send them messages, etc.)? I would prefer open-source possibilities, if possible.
For seeing what's going on with all those devices, and notification of problems, and automating any processes, then DevicePilot is worth a look. It's not an IoT platform though (there are plenty of those).
Cumulocity offers a complete Software-as-a-service to do that.
Raspberry Pi integration is described here:
http://www.cumulocity.com/guides/devices/raspberry-pi/ (Java Agent)
or here:
http://www.cumulocity.com/guides/devices/janztec/ (C++ Agent)
The agents are open source (under https://bitbucket.org/m2m/ ), the server-side platform is not open source (except examples). But you can get a free trial (see http://www.cumulocity.com/ ).
On
It would work like that: upon start, raspberry would register to the service, once done, the service allow me to check which machines are up / down, send them messages etc.
This is supported out of the box without programming. Please see links above.
Stefan (working for Cumulocity)
Microsoft Live Custom Domains are now shutting down. I have been using this service for e-mail hosting for all my Microsoft Azure Websites, but now it is gone and no viable replacement is in sight. Do you have some idea what alternative approach exist for hosting multiple mailboxes for multiple websites hosted on Azure?
Your cheapest option is to have one (single) google apps account which will cost you around $5 per month. Make it something very general like mail#yourdomain.com. Then in the google apps dashboard make it a catch all address. This will make the single inbox catch all email for accounts such as Sales#, support# bob# or whatever at your domain.
Then you can set Gmail filters to sort (label) each incomming message based on who it was sent to. For example you can have messages sent to frank# automatically labeled as frank.
Next you want to create regular old gmail accounts for all of your individual users. I am going to follow the example of creating a box for frank#company.com for this instance.
Create Google Apps primary account (catch all) as mail#company.com
Create Gmail account for frank.company#gmail.com (regular gmail account)
Create filter rule on Google Apps account for all messages sent to frank#company.com to be forwarded to frank.company#gmail.com. You can further mark them as read or delete them upon forward.
In the frank.company#gmail.com create a sending alias as frank#company.com. Google will give you a 4 digit code, and now when logging in as frank.company#gmail.com i can both send as frank#company.com and recieve all email since its forwarded to this account.
Also make sure to set default reply:to addresses in case you send from the frank.company#gmail.com address.
Using the technique above you can get all the benefits of having a pro google apps account (dkim, spf, 25gb inbox) and with a little bit of configuration you can setup multiple gmail accounts which run off the single account. We use the technique above and it works flawlessly. The only thing that doesn't work is mailbox delegation, which is not that great.
If you wanted to save the $5 you could get away with using something like GoDaddy free email forwarding, but then you would be limited to godaddys 250 message limit per day.
The approach above just works.
I feel your pain. Had to have some tough talks with many of my customers when the free Google Apps option was discontinued.
I found two routes:
Find a hosted Exchange type solution. This has the advantages of any hosted solution. It is managed for you. You can get started with around 50 USD / user / year and services are provided by the likes of Microsoft, Google and Rackspace, like stated in the other answers.
(Which is the route I chose) Host your own Exchange server on AWS EC2 or Azure. Thanks to Microsoft License Mobility, you can install an Exchange license on a cloud server and provide email addresses for your customers` domains yourself. This will allow you to share the cost of the Exchange license between all your customers and if you reach the critical mass, this can save a lot compared to the pay-per-user-per-month models for most hosted solutions.
I am stil looking for a free alternative, but have yet to find one that can match the features that were available in the free version of Google Apps.
EDIT: I was thinking about this again last night and came up with another idea. I am not a Linux guy, so I would not be happy to do this for production mail server. For someone who is "bilingual" (i.e. ok with both MS and Linux solutions) or of a more adventurous nature than me, could take route 2 with a linux server and an open source mail server solution. I am sure this will lower the cost even more significantly, since you will not need to pay for the mail server licence and also per-hour instance costs for Linux servers are lower. This might even create a whole new revenue stream.
Zoho provides a Google Apps like deal for 5 users for free:
https://www.zoho.com/mail/zohomail-pricing.html
Up to five users
5GB/User
25MB attachment limit
Web access only
Email hosting for single domain
I just finished installing a mail server in Azure in a Linux virtual machine. So far seems ok.
The total cost of operation is about 10€ a month since neither Ubuntu (the OS) nor iredMail (the mail server) nor Postgres (the database) have any licensing fees.
Regarding the block on Azure IPs I do believe that most users saying that did not correctly configured their servers. And by that I mean that they didn't configured the PTR reverse DNS on Azure, which allows other mail servers to check if that IP is allowed to send and receive mail from that domain.
Also make sure you add the SPF DNS entry for your mail server. You can't blame a mail server to blacklist you if you don't minimize the risks of SPAM.
Hope this helped you.
Useful links:
IredMail Server - http://www.iredmail.org/
Reverse DNS in Azure - http://azure.microsoft.com/blog/2014/07/21/announcing-reverse-dns-for-azure-cloud-services/
First of all you need to identify how much and what all services do you need?
If it is just an IMAP/POP3 Email Box, then best option is a Virtual Server or Virtual Machine with cpanel, once installed with daily backup runs good for years, you get unlimited email accounts and unlimited space !!! You can increase your VM dynamically up when you need it. Drawback is, it takes little maintenance once in a while. But most likely cpanel auto update is very stable and I have VPS running for 5 years and every year we are just increasing our disk space.
If you want calendar along with live docs editing etc, then you have to go with Google Apps which is cheaper then MS Exchange. But if you need strict Exchange kind of services, then you will have to go with hosted Exchange.
I will not recommend spending money for Rackspace or any such Cloud Email which is priced per user, which is total waste of money as they do not offer anything apart from linux server with cpanel. Those services are only for non IT people. Since you have already asked question on Stack Overflow, you can easily setup and manage cpanel based linux OS.
you could run a Ubuntu VM in Azure and set up Postfix
You can install a Free Mail Server on a Virtual Machine on Azure like:
https://www.hmailserver.com/download
I have found the same problem myself.
The only alternative in the past would have been to use Google's equivalent service, but they have also stopped it.
Realistically, there isn't a free answer to using custom domains with emails that I am aware of. Both Microsoft and Google offer paid for services, but cost per user/mailbox, per-year - compared to their free services this is a big jump in price.
Google charge £33 per user/mailbox, per year.
Microsoft are slightly dearer at £39 per user/mailbox, per year, but include access to online versions of Office for each user too.
For my situation, the Microsoft route may be the better option, based on my customer preferences but I am sure that the Google service is equally adequate.
Hope this helps. (But let me know if you find a better alternative!!!)
How many real people do you have reading e-mail? As many as you got mailboxes? If not, then I really suggest you go for Exchange Online from Microsoft which goes for $4 per user (not mailbox) per month.
The trick is, that once you've set up your domains, you create a Shared Mailbox through PowerShell and while doing so, you give the real people (you pay for) the rights to read and send as. The cool thing is, that the user does not need to do anything. The mailbox simply appears in their Outlook.
I'm trying to study about high frequency trading systems. Whats the mechanism that HFT use to connect with the exchange and whats the procedure (does it has to go through a broker or is it direct access, if it's direct access what sort of connection information that i require)
Thanks in advance for your answers.
Understand that there are two different "connections" in an HFT engine. The first is the connection to a market data source. The second is to a clearing resource. As mentioned in kpavlov's answer, a very expensive COLO (co-location) is needed to get as close to the data source/target as possible. Depending on their nominal latency these COLO resources cost thousands of dollars per month.
With both connections, your trading engine must be certified by the provider (ICE, CME, etc) to comply with their requirements. With CME the certification process is automated, with ICE it employs human review. In any case, the certification requires that your software demonstrate conformance to standards and freedom from undesirable network side effects.
You must also subscribe to your data source(s) and clearing service, neither is inexpensive and pricing varies over a pretty wide range. During the subscription process you'll gain access to the service providers technical data specification(s)-- a critical part of designing your trading engine. Using old data that you find on the Internet for design purposes is a recipe for problems later. Subscription also gets you access to the provider(s) test sites. It is on these test sites that you test and debug your engine.
After you think you engine is ready for deployment you begin connecting to the data/clearing production servers. This connection will get you into a place of shadows-- port roulette. Not every port at the provider's network edge has the same latency. Here you'll learn that you can have the shortest latency yet seldom have orders filled first. Traditional load balancing does little to help this and CME has begun deployment of FPGA-based systems to ensure correct temporal sequencing of inbound orders, but it's still early in its deployment process.
Once you're running you then get to learn that mistakes can be very expensive. If you place an order prior to a market pre-open event the order is automatically rejected. Do it too often and the clearing provider will charge you a very stiff penalty. Other things can also get you penalized or even kicked-off the service if your systems are determined to be implementing strategies to block others from access, etc.
All the major exchanges web sites have links to public data and educational resources to help decide if HFT is "for you" and how to go about it.
It usually requires an approval from exchange to grant access from outside. They protect their servers by firewalls so your server/network need to be authorized to access.
Special certification procedure with technician (by phone) is usually required before they authorize you.
Most liquidity providers use FIX protocol or custom APIs. You may consider starting implementing your connector with QuickFix, but it may become a bottleneck later, when your traffic will grow.
Information you need to access by FIX is:
Server IP
Server port
FIX protocol credentials:
SenderCompID
TargetCompID
Username
Password
Other fields
I am developing a customer account system for a chain of recycling centers in the Northwest US. One of our key features is that our customers can set up accounts that are credited with their bottle deposit refunds, instead of always disbursing cash. Customers can also drop off bags of recyclables that are processed on-site and credited. Each center runs near capacity and can physically process cans and bottles when offline, so we don't have a lot of leeway for IT infrastructure to shut down everything when the Internet goes out.
Basically, I've been asked to develop a customer account system that will allow credits from a retail center to be posted to accounts, even if telecommunications with our central server breaks down for a period of hours. This will allow the center to keep processing and crediting customers when the pipes get clogged. Certain transactions, like withdraws, do NOT need to occur in this situation, since we can't accurately get the customer's current balance.
We are a 100% Windows shop, and the IT manager and network admin don't want to get near anything *nix. Each retail center has an on-premise dedicated Windows Server, so that seems like a logical place to start.
I'm a huge fan of ServiceStack, and the REST-ful message-based paradigm seems like might work. I'd create a "Credit" message and send it to the local server. A message broker there would log the request and attempt to forward that message to the central server where it is processed. In case the central server were down, I would rely on the MQ's reliable messaging protocol to hold on to it until telecommunications are restored. The overall anticipated volume is 100s to low 1,000s of messages out of each center, so low by modern computing terms.
The Redis MQ Client / Server for ServiceStack looks interesting, but since the Windows Redis server is explicitly labeled "prototype" and "not production quality", there is a 0% chance of being able to leverage it.
So, ultimately the questions are:
Is a reliable messaging system the right type of solution for this problem? Are there other approaches I should consider?
Are there alternatives to Redis that play well with ServiceStack? Is there a "production quality" NoSQL server replacement I can use on Windows?
I've looked briefly at RabbitMQ. Might that be an option? My Googling doesn't show any active integration between it and ServiceStack, so I'm leery of writing something from the ground up.
Ideally the overhead of my solution is low enough we can perform a synchronous update and return a "current balance" receipt to a customer if everything is working well. Is this a realistic?
A production solution for running Redis on windows is to run redis-server inside a Linux VM on windows with Vagrant.
There is current a feature request to add more MQ Options to ServiceStack. Rabbit MQ is expected to be the next MQ adapter to be supported in future.
As a follow-up, MS Open Tech has released a "production-ready" native implementation of Redis 2.8.9. GitHub link.