Warning : I'm new in application architecture and officially speaking - this is the first time I'm designing something this big. This is also my own application so I got full authority to change things.
I'm building a serverless application which consists of an on-demand application streaming platform.
Customers who seek to try a specific application (usually large and expensive ones like Photoshop or Solidworks for example) could have the possibility to directly try one from their computers, on their browser, while the application is running on a similar to their computer type of infrastructure.
I'd use the CI/CD pipelines and IaC technology to build EC2 infrastructure that will host these applications and use those same technologies to destroy that infrastructure, since it's volatile.
So to create/destroy that EC2 infrastructure I use the GitLab API.
I've thus decided to go with AWS Lambda & GitLab for now.
Now the architecture questions :
Is it better to have one serverless function that handles everything or several functions ?
I'm planning to destroy the EC2 infrastructure after a certain amount of time (10-15 minutes). How should I schedule HTTP communication? Should I use a queue like SQS? Should I use some database and check every minute?
Again, thanks a lot for your wisdom!
Edit : Clarification on some stuff.
Related
In Google's latest docs, they say to test Go 1.12+ apps locally, one should just go build.
However, this doesn't take into account all the routing etc that would happen in the app engine utilizing the app.yaml config file.
I see that the dev_appserver.py is still included in the sdk. But it doesn't seem to work in Windows 10.
How does one test their Go App Engine App locally with the app.yaml. ie: as an actual emulated app engine app.
Thank you!
On one hand, if your application consists of just the default service I would recommend to follow #cerise-limón comment suggestion. In general, it is recommended for the routing logic of the application to be handled within the code. Although I'm not a Go programmer, for single service applications that use static_files and static_dir there shouldn't be any problems when testing the application locally. You might also deploy the new version without promoting traffic to it in order to test it as explained here.
On the other hand, if your application is distributed across multiple services and the routing is managed through the dispatch.yaml configuration file you might follow two approaches:
Test each service locally one by one. This could be the way to go if each service has a single responsibility/functionality that could be tested in isolation from the other services. In fact, with this kind of architecture the testing procedure would be more or less the same as for single service applications.
Run all services locally at once and build your own routing layer. This option would allow to test applications where services needs to reach one another in order to fulfill the requests made to them.
Another approach that is widely used is to have a separate project for development purposes where you could just deploy the application and observe it's behavior in the App Engine environment. As for applications with highly coupled services it would be the easiest option. But it largely depends on your budget.
I have some hardware devices that send some data that need to be stored on the cloud server and also I need to do some real time processing on them.
The data they send need to be preserved for months in some custom binary files. These files related to each device can grow in size up to 10GB over time.
There will client programs (mobile / web) that will be looking at the processed data at real time.
My prefered choice of language is C/C++/C#, since there is time sensitive number crunching involved.
Goal is write scalable application that can have thousands of such devices monitored on the cloud.
Do I have to upfront write the code for running on the cloud ( undestand Azure / amazon EC2) ? Can I write multi threaded desktop application and later migrate to cloud ?
I have used Message passing interface (MPI) in the past for clusters. Can I still use MPI ?
If I use microsoft azure API can I still host my software on Amazon cloud ?
For mobile devices to talk to the server, I understand that I need to have a webservice running. how can I convert a desktop program writeen in C++ / C# to act as a web service talking to client?
Are there any 3rd part frame works or tools taht can help me with my work ?
With most cloud compute services you can deploy an off-the-shelf server and install your own software on it. So, yes, you can write and test you application locally then migrate to the cloud once you get all the bugs worked out. Here are the available EC2 server configurations.
I have not tried MPI but you should be able to run just about anything you want on the servers in the cloud. However, Amazon does offer the Simple Queue Service which provides message passing in the cloud. Your software does not need to run in the cloud to use this service.
I have not used Azure. I doubt there are any restrictions regarding which external servers you use for storage and/or compute. However, keeping your cloud storage and compute resources within a single provider will reduce costs, improve performance and provide you with a unified management interface and billing system.
Web servers are fairly simple things. See this post. That took me about 10 seconds to find.
There is plenty of third party software out there. Figure out what you need in more detail and ask more specific questions
I have a project and I'm planning to start the web app as an Azure Web Site and then migrate it to an Azure Cloud Service (also called Hosted Service) if it is needed as a scale strategy.
The decision is because I read that Azure Web Sites are more simple and fast to develop with almost no Azure-specific configurations or code. So starting fast and simple is a good starting point for the project.
But, is that a good starting point for you?
Is migrating an Azure Web Site to an Azure Cloud Service the same as you were migrating a normal ASP.NET Website to an Azure Cloud Service?
Would you start with an Azure Cloud Service right from the beginning? If yes, why?
Thanks for your time.
There are benefits to both deployment models, it will eventually come down to what you are trying to achieve and ultimately the success of your application.
Below I've outlined the Pros and Cons of each of the models to ensure that you're making the right choice for your applications goals.
Windows Azure Web Sites
You have properly identified that Windows Azure Web Sites is a great starting point for an application, however you could also consider that Web Sites does offer enough scalability for many solutions.
Pros
10 Free sites during preview [Free for 12 months]
Easy Deployment (use Git, TFS, Web Deploy or FTP)
Quick Scalability (You can move to your own dedicated cluster [aka reserved standard])
Simple Development (Supports Classic ASP, ASP.NET, Node.js, Python & PHP)
Persistent Environment (most people are used to this)
Cons
No SSL Support on Custom Domains
in Preview (currently no SLA)
Windows Azure Cloud Services
Cloud Services (formerly known as Hosted Services) is definitely the vision for the future of Web Applications. It is built with resiliency in mind to keep the cost of applications affordable by scaling to meet demand, and dial back capacity when your traffic slows.
Pros
Increased control over the cost of your application (if architected correctly)
Flexibility (You have full control over the environment)
SSL Support
Language Agnostic
Web Server Agnostic (although IIS is available by default)
Auto Management of Servers
Cons
Architecture should be carefully considered
Deployment time is slower (Slows development cycle)
Things to consider for Portability
The items above might have given you enough to plan the immediate future of the application and it is very likely that you might want to consider Cloud Services in the future (it fits a number of application scenarios better in the long run).
Here is a list of things to help portability between Web Sites to Cloud Services:
Start thinking Stateless
Windows Azure Web Sites is nice as it is a persistent environment, which means you are able to store things like session state and assets to the disk.
Although this is a good feature, it's best to start planning towards a stateless application, if your end goal is to be in Cloud Services. Here are a few things you can do to start thinking stateless:
Don't rely on Session State
If you need it, come up with a strategy to make it scale (Caching Service, SQL, or Storage)
Use the Storage Service
Assets such as Static HTML, css, javascript and images are better placed in Storage
Avoids additional bandwidth on your Web Site (potentially stay shared longer for lower cost)
Can be CDN Enabled, provides a better experience for International markets
Easier to update web assets when application is migrated to Cloud Services
Storing User content
If your application already stores to the Storage Service, there is one less code modification in the future when moving to cloud services.
Make it easy to discover patterns in your Data
The benefit of Cloud Services is it enables you to reduce cost by only scaling what needs scaled. Starting the process of identifying your scale units i.e. How you partition your database or Tables in Storage.
I read all post and all of them are very helpful.
In addition to all post , I found an info on msdn : Windows Azure Websites, Cloud Services, and VMs: When to use which?
With Windows Azure Websites you can:
Build highly scalable web sites on Windows Azure.
Quickly and easily deploy sites to a highly scalable cloud environment that allows you to start small and scale as needed.
Use the languages and open source applications of your choice then deploy with FTP, Git or TFS, and easily integrate Windows Azure services like SQL Database, Caching, CDN and Storage.
With Cloud Services you can:
Build or extend your enterprise applications on Windows Azure.
Create highly-available, scalable applications and services using a rich PaaS environment. Support advanced multi-tier scenarios, automated deployments and elastic scale. Deliver great SaaS solutions to customers anywhere around the world.
And also there is summarizes the option on msdn :
And comparing some features Web Sites and Cloud Services on msdn:
Azure is a great place to have your app, but there are some considerations you need to know before start migrating it.
Azure Websites and Hosted Services are really trivial to deploy. With
Visual studio you generate the package and simply upload it. Then you
have a Development environment to check it. If it's ok for you, swap
ips. If it's not ok for you, upgrade again.
Your instances have some properties that could be annoying. For
example, you cannot be sure about your IP. Then if your app works
with some provider using IP restriction, you will need to figure out
how to proceed.
More considerations. Your "server" could be reimaged at any moment.
If you store something on the local disc, that file could go away at any moment.
Azure works very nice if you have at least 2 instances or more for
each website. Maybe your app is not prepared for that. The first step
will be managing the sessions with the appFabric. Is really
easy, just a change on your web config. Be careful because this
session state doesn't work exactly as the "old one". You cannot store
non-serializable objects (should be easy to adapt) or a very large objects (more than 8MB).
If you are going to develop something from zero, I suggest you to start into azure from the beginning. The reason is simple: it's really cheap to start and you will not pay serious money until the app have lot's of visits. It's also very cheap to setup a SQLAzure and a storage account. One you have all in place, it's easy to add more instances or scale up.
Example:
Imagine you have an idea and you wish to show up to some possible investors.
You start setting up a little SQLAzure database (1GB ), $9,99 monthly.
Then you build a site and you put 2 extra small instances, $18,72 monthly.
Let's say you need 100 GB of space (images, backups, ...), $12,50 monthly.
At his point, you have all in place to start your business paying less than $50 monthly.
If you site have exit and the visits starts to come, you change your instances for small instances (it's really dangerous to have production environment with extra small instances, because do not have cpu reservation). Then you change the extra small cost ($18,71) up to $57,60. Maybe you need more space to that SQL Azure? etc...
prices calculated from here: http://www.windowsazure.com/en-us/pricing/calculator/?scenario=web .
Those are few tips, there is a lot more. My advice is to start a trial account and play with it.
Final advice: Its very easy to solve everything just purchasing more resources. Sometimes you need to refactor and optimize your code. If you simply add more resources each time you have a problem, you could end with a huge bill and a very messy code.
Hope it helps!
Another advantage of Windows Azure Cloud Services over Web Sites is that a cloud service can be added to an Azure Virtual Network. This can give it access to on-premises resources like databases. So if your requirements are such that you need the scalability offered by Azure but need to keep your data on-premises due to security restrictions, cloud services is a better choice.
Azure web sites cannot be part of an Azure virtual network. To access on-premises resources mechanisms such as Azure Service Bus Relay must be configured.
We've had our web site running on PHP on some hosting and at some point decided to move it to Azure (where sits main part of our service). We've started with Azure Web Sites which was great from development point of view (mainly integration with git). But after about a week of testing (when we've decided to actually move the production web site) we've found that currently
No SSL for custom domains
Custom domains are available only for reserved instances (no shared infrastructure)
SLA
So we moved to Hosted Service. The main problem for us was lack of ability of simple deployment (need to build package and upload whole package of the web site), and found solution was to use dropbox - as a startup task for role, we're installing dropbox service on the machine, which takes all the web site from dropbox, which in turn have SVN checked out folder, so site updates became very easy.
I just started learning Ruby on rails and I was wondering what Heroku really is? I know that its a cloud that helps us to avoid using servers? When do we actually use it?
Heroku is a cloud platform as a service. That means you do not have to worry about infrastructure; you just focus on your application.
In addition to what Jonny said, there are a few features of Heroku:
Instant Deployment with Git push - build of your application is performed by Heroku using your build scripts
Plenty of Add-on resources (applications, databases etc.)
Processes scaling - independent scaling for each component of your app without affecting functionality and performance
Isolation - each process (aka dyno) is completely isolated from each other
Full Logging and Visibility - easy access to all logging output from every component of your app and each process (dyno)
Heroku provides very well written tutorial which allows you to start in minutes. Also they provide first 750 computation hours free of charge which means you can have one processes (aka Dyno) at no cost. Also performance is very good e.g. simple web application written in node.js can handle around 60 - 70 requests per second.
Heroku competitors are:
OpenShift by Red Hat
Windows Azure
Amazon Web Services
Google App Engine
VMware
HP Cloud Services
Force.com
It's a cloud-based, scalable server solution that allows you to easily manage the deployment of your Rails (or other) applications provided you subscribe to a number of conventions (e.g. Postgres as the database, no writing to the filesystem).
Thus you can easily scale as your application grows by bettering your database and increasing the number of dynos (Rails instances) and workers.
It doesn't help you avoid using servers, you will need some understanding of server management to effectively debug problems with your platform/app combination. However, while it is comparatively expensive (i.e. per instance when compared to renting a slice on Slicehost or something), there is a free account and it's a rough trade off between whether it's more cost effective to pay someone to build your own solution or take the extra expense.
Heroku Basically provides with webspace to upload your app
If you are uploading a Rails app then you can follow this tutorial
https://github.com/mrkushjain/herokuapp
As I see it, it is a scalable administrated web hosting service, ready to grow in any sense so you don't have to worry about it.
It's not useful for a normal PHP web application, because there are plenty of web hosting services with ftp over there for a simple web without scalability needs, but if you need something bigger Heroku or something similar is what you need.
It is exposed as a service via a command line tool so you can write scripts to automate your deployments. Anyway it is pretty similar to other web hosting services with Git enabled, but Heroku makes it simpler.
That's its thing, to make the administration stuff simpler to you, so it saves you time. But I'm not sure, as I'm just starting with it!
A nice introduction of how it works in the official documentation is:
https://devcenter.heroku.com/articles/how-heroku-works
Per DZone: https://dzone.com/articles/heroku-or-amazon-web-services-which-is-best-for-your-startup
Heroku is a Platform as a Service (PaaS) product based on AWS, and is vastly different from Elastic Compute Cloud. It’s very important to differentiate ‘Infrastructure as a Service’ and ‘Platform as a Service’ solutions as we consider deploying and supporting our application using these two solutions.
Heroku is way simpler to use than AWS Elastic Compute Cloud. Perhaps it’s even too simple. But there’s a good reason for this simplicity. The Heroku platform equips us with a ready runtime environment and application servers. Plus, we benefit from seamless integration with various development instruments, a pre-installed operating system, and redundant servers.
Therefore, with Heroku, we don’t need to think about infrastructure management, unlike with AWS EC2. We only need to choose a subscription plan and change our plan when necessary.
That article does a good job explaining the differences between Heroku and AWS but it looks like you can choose other iaas (infrastructure) providers other than AWS. So ultimately Heroku seems to just simplify the process of using a cloud provider but at a cost.
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 8 years ago.
Improve this question
I have thought a lot recently about the different hosting types that are available out there. We can get pretty decent latency (average) from an EC2 instance in Europe (we're situated in Sweden) and the cost is pretty good. Obviously, the possibility of scaling up and down instances is amazing for us that's in a really expansive phase right now.
From a logical perspective, I also believe that Amazon probably can provide better availability and stability than most hosting companies on the market. Probably it will also outweigh the need of having a phone number to dial when we wonder anything and force us to google the things by ourselves :)
So, what should we be concerned about if we were about to run our web server on EC2? What are the pro's and cons?
To clarify, we will run a pretty standard LAMP configuration with memcached added probably.
Thanks
So, what should we be concerned about if we were about to run our web server on EC2? What are the pro's and cons?
The pros and cons of EC2 are somewhat dependent on your business. Below is a list of issues that I believe affect large organizations:
Separation of duties Your existing company probably has separate networking and server operations teams. With EC2 it may be difficult to separate these concerns. ie. The guy defining your Security Groups (firewall) is probably the same person who can spin up servers.
Home access to your servers Corporate environments are usually administered on-premise or through a Virtual Private Network (VPN) with two-factor authentication. Administrators with access to your EC2 control panel can likely make changes to your environment from home. Note further that your EC2 access keys/accounts may remain available to people who leave or get fired from your company, making home access an even bigger problem...
Difficulty in validating security Some security controls may inadvertently become weak. Within your premises you can be 99% certain that all servers are behind a firewall that restricts any admin access from outside your premises. When you're in the cloud it's a lot more difficult to ensure such controls are in place for all your systems.
Appliances and specialized tools do not go in the cloud Specialized tools cannot go into the cloud. This may impact your security posture. For example, you may have some sort of network intrusion detection appliances sitting in front of on-premise servers, and you will not be able to move these into the cloud.
Legislation and Regulations I am not sure about regulations in your country, but you should be aware of cross-border issues. For example, running European systems on American EC2 soil may open your up to Patriot Act regulations. If you're dealing with credit card numbers or personally identifiable information then you may also have various issues to deal with if infrastructure is outside of your organization.
Organizational processes Who has access to EC2 and what can they do? Can someone spin up an Extra Large machine and install their own software? (Side note: Our company http://LabSlice.com actually adds policies to stop this from happening). How do you backup and restore data? Will you start replicating processes within your company simply because you've got a separate cloud infrastructure?
Auditing challenges Any auditing activities that you normally undertake may be complicated if data is in the cloud. A good example is PCI -- Can you actually always prove data is within your control if it's hosted outside of your environment somewhere in the ether?
Public/private connectivity is a challenge Do you ever need to mix data between your public and private environments? It can become a challenge to send data between these two environments, and to do so securely.
Monitoring and logging You will likely have central systems monitoring your internal environment and collecting logs from your servers. Will you be able to achieve the monitoring and log collection activities if you run servers off-premise?
Penetration testing Some companies run periodic penetration testing activities directly on public infrastructure. I may be mistaken, but I think that running pen testing against Amazon infrastructure is against their contract (which make sense, as they would only see public hacking activity against infrastructure they own).
I believe that EC2 is definitely a good idea for small/medium businesses. They are rarely encumbered by the above issues, and usually Amazon can offer better services than an SMB could achieve themselves. For large organizations EC2 can obviously raise some concerns and issues that are not easily dealt with.
Simon # http://blog.LabSlice.com
The main negative is that you are fully responsible for ALL server administration. Such as : Security patches, Firewall, Backup, server configuration and optimization.
Amazon will not provide you with any OS or higher level support.
If you would be FULLY comfortable running your own hardware then it can be a great cost savings.
i work in a company and we are hosting with amazon ec2, we are running one high cpu instance and two small instances.
i won't say amazon ec2 is good or bad but just will give you a list of experiences of time
reliability: bad. they have a lot of outages. only segments mostly but yeah...
cost: expensive. its cloud computing and not server hosting! a friend works in a company and they do complex calculations that every day have to be finished at a certain time sharp and the calculation time depends on the amount of data they get... they run some servers themselves and if it gets scarce, they kick in a bunch of ec2's.
thats the perfect use case but if you run a server 24/7 anways, you are better of with a dedicated rootserver
a dedicated root server will give you as well better performance. e.g. disk reads will be faster as it has a local disk!
traffic is expensive too
support: good and fast and flexible, thats definately very ok.
we had a big launch of a product and had a lot of press stuff going on and there were problems with the reverse dns for email sending. the amazon guys got them set up all ripe conform and nice in not time.
amazon s3 hosting service is nice too, if you need it
in europe i would suggest going for a german hosting provider, they have very good connectivity as well.
for example here:
http://www.hetzner.de/de/hosting/produkte_rootserver/eq4/
http://www.ovh.de/produkte/superplan_mini.xml
http://www.server4you.de/root-server/server-details.php?products=0
http://www.hosteurope.de/produkt/Dedicated-Server-Linux-L
http://www.klein-edv.de/rootserver.php
i have hosted with all of them and made good experiences. the best was definately hosteurope, but they are a bit more expensive.
i ran a CDN and had like 40 servers for two years there and never experienced ANY outage on ANY of them.
amazon had 3 outages in the last two months on our segments.
One minus that forced me to move away from Amazon EC2:
spamhaus.org lists whole Amazon EC2 block on the Policy Block List (PBL)
This means that all mail servers using spamhaus.org will report "blocked using zen.dnsbl" in your /var/log/mail.info when sending email.
The server I run uses email to register and reset passwords for users; this does not work any more.
Read more about it at Spamhaus: http://www.spamhaus.org/pbl/query/PBL361340
Summary: Need to send email? Do not use Amazon EC2.
The other con no one has mentioned:
With a stock EC2 server, if an instance goes down, it "goes away." Any information on the local disk is gone, and gone forever. You have the added responsibility of ensuring that any information you want to survive a server restart is persisted off of the EC2 instance (into S3, RDS, EBS, or some other off-server service).
I haven't tried Amazon EC2 in production, but I understand the appeal of it. My main issue with EC2 is that while it does provide a great and affordable way to move all the blinking lights in your server room to the cloud, they don't provide you with a higher level architecture to scale your application as demand increases. That is all left to you to figure out on your own.
This is not an issue for more experienced shops that can maintain all the needed infrastructure by themselves, but I think smaller shops are better served by something more along the lines of Microsoft's Azure or Google's AppEngine: Platforms that enforce constraints on your architecture in return for one-click scalability when you need it.
And I think the importance of quality support cannot be underestimated. Look at the BitBucket blog. It seems that for a while there every other post was about the downtime they had and the long hours it took for Amazon to get back to them with a resolution to their issues.
Compare that to Github, which uses the Rackspace cloud hosting service. I don't use Github, but I understand that they also have their share of downtime. Yet it doesn't seem that any of that downtime is attributed to Rackspace's slow customer support.
Two big pluses come to mind:
1) Cost - With Amazon EC2 you only pay for what you use and the prices are hard to beat. Being able to scale up quickly to meet demands and then later scale down and "return" the unneeded capacity is a huge win depending on your needs / use case.
2) Integration with other Amazon web services - this advantage is often overlooked. Having integration with Amazon SimpleDB or Amazon Relational Data Store means that your data can live separate from the computing power that EC2 provides. This is a huge win that sets EC2 apart from others.
Amazon cloud monitoring service and support is charged extra - the first one is quite useful and you should consider that and the second one too if your app is mission critical.