Lately, I have been struggling to understand what is my network speed (downlink) between nodes on AWS (in a multi-homed cluster, computers in different regions).
I have a lot of fluctuations when I measure it with a script which I have written (based on this link and SCP) or with Iperf.
I believe it is based on network use which changes rapidly (mostly between regions), but I still don't understand AWS documentation about what is the performance I am paying for, a minimum and a maximum downlink rate for example (aws instances).
At first, I have tried the T2 type, and as I saw it had burst CPU performance, I thought that maybe the NIC performance is also bursty so I have moved to M4 type, but I have got the same problems with M4.
Is there any way to know my NIC downlink rate based on the type and flavor?
*I have asked a similar question on the AWS forum, but I haven't got a response (https://forums.aws.amazon.com/thread.jspa?threadID=296389).
There is no way to get a better indication that your measuring. AWS does not publish anything indicating this performance, and unless we are talking the larger instance where network performance is actually specifically given. I.e. m5.12xlarge having 10 gbps. Most likely network performance does have a burst component for smaller instance types.
There are pages with other peoples benchmarks, but you won't find any official answer for any of this.
I created an Amazon EC2 Auto Scaling group, where it should have at least 1 server all the time.
Add up 2 servers when CPU utilization passes beyond 80%
Terminate 2 servers when CPU utilization comes down less then 30%
Challenge here is, How should I increase/decrease CPU utilization? I cannot connect to any instance or use CLI since I am in Office system / restricted AWS access.
Is there a way to test this despite of these restrictions?
There is a way to stress test an instance or container (assuming it is Linux based) using Stress, a package that is designed to crank up the CPU for a specified amount of time and then bring the CPU percentages down after the specified amount of time. It has other parameters to customize the testing.
My personal favorite tool for testing system response and DR is to use Netflix's ChaosMonkey. It is an open source project, backed by Netflix that is designed to test fault tolerance. Using it in production comes down to personal preference, but it is a tool for testing systems.
If you want to test the "real" situation, then you will need a way to generate load on the system.
This could be artificial load (eg triggering a program that does calculations, just to spin the CPU) or a real-world simulation of actual activities that you system will perform.
There is no need to test whether Amazon EC2's Auto Scaling actually works — there would be issues shown on the AWS Status Page if that were the case — so I presume you just wish to test your own configuration. In this case, you should really be testing a real world scenario, such as simulating a quantity of simultaneous users doing typical activities that users would perform.
If you do any other form of testing (such as fake increasing of CPU load), you're not really testing the real situation in which you want Auto Scaling to perform, so the results of your test won't actually be useful.
For example, it might be that your application runs into memory issues or single-threading issues way before it hits any CPU limits. That would be something you'd really like to know before throwing real users at your system.
You'll have to excuse my ignorance on this one...but honestly, I've had a hard time finding clarity on this. That being said, I'm looking for a non technical answer...something in layman's terms!
Anyways, I've been playing around building a web app (first time obviously) and I'm getting to the point where I've started looking into hosting services. A quick google search and a few blogs later, I thought AWS would be a good place to start, since they give a free-year trial. I don't care about speedy upstarts or other hosting serves, so save your key strokes on offering other services.
My question is based on the fact that AWS charges "Linux Usage per hour" and they also use this term "instance". Yeah...an "instance" is an "object", which is also above my head (probably the real source of the problem), but that was the extent I was able to learn via a google search. That being said, I don't know how to translate the cost into a ball park cost. Yes, I can probably use the trial to help monitor predictable costs, but I don't want to go through the effort of learning one hosting companies system just to find out it's not going to work in the end.
OK...so hopefully by now you see where I'm coming from. What is an "instance" and how do I use the "Linux Usage per hour" to estimate cost? Is an instance a server? For example if I start NGINX is that in instance? Is it just one instance running NGINX or does every VPN represent an instance? If I have 100 people calling the server at once, can they fit on one instance? If I start another server say, Apache or Node, does that become another instance? If I connect to a database, is that an instance? Do instances start as needed? Yes, I know, that's more than one question...I'm just trying to express my confusion.
If I'm suppose to choose a pricing model from this list, "Linux Usage per hour", I need to know what them mean by "Linux Usage". If it's based on an "instance", I need to know what that is. So please, in layman's terms, help clear this up. Maybe some examples or analogies, but no deep technical stuff.
This is more a side note, but I was reading this article and it said
For a client needing to run 800 virtual instances, the annual cost of
a private cloud came to below $400,000 vs. somewhere between $800,000
and $1.2 million for public cloud services.
Considering I don't know what an instance is, that kinda made me a bit nervous...WAAAAAAyyyyyy outta my price range! Yes, it's obviously a big company, but can you imagine "hitting the lottery" with an app everyone loves then before you know it, AWS hits you with a bill of $1,000,000. Or even worse, your security sucks and someone spawns millions of these "instances"...help alive my paranoia!!
Basically, an instance is a virtual machine, which looks very much like a server. As such it's running an operating system - e.g. linux - which is capable of running many programs (aka 'processes' or sometimes, 'services') at the same time.
To go through your questions (some of the explanations below are not technically accurate, but are hopefully more explanatory for it - if anything is obvious or already known, apologies - trying not to assume any knowledge)
An instance is an object
This definition is coming up in your searches because 'instance' has many definitions in different situations. If you see the definition of 'instance' as an object, it's from the topic of object oriented programming languages - you define a class in your code (kind of like a 'template'), and then create instances of the class - kind of like real copies of the template.
Amazon borrowed the term to be analogous - because in the 'cloud' world, you can create an AMI (Amazon Machine Image - the template) and then create lots of instances that are copies or clones of that template.
Is an instance a server?
In terms of what you can do with it, yes, it's a server.
(Technically it's a virtual server - Amazon runs multiple virtual servers on each physical server.)
how do I use the "Linux Usage per hour" to estimate cost?
Estimate how long you will have your instance running for in hours per month, multiply it by cost per hour and you will have your estimated cost per instance per month.
e.g. - one instance always turned on would be - 24 hrs * 31 days = 744 hours. At $0.013/hr (for a t2.micro) that would be 744 * $0.013 = $9.672/mth.
(And that's the reason the free tier gives you 750 hours of instance time per month.)
Instances come in different types and sizes and each size costs a different amount. If you are not sure what size you need, I'd start with the smallest until you discover you need more - which would be when your program starts running too slowly.
For example if I start NGINX is that in instance?
Nginx is a program that runs as a daemon in linux terms - a program that runs in the background so it's always on. It will be one of the many programs running on the server (aka the instance)
If I have 100 people calling the server at once, can they fit on one instance?
It depends - on how big your instance is, and how efficient the program is that is responding to their requests. If you are just getting started learning to program websites, I wouldn't worry about handling 100 people issuing requests to the server all at once just yet - walk before you run :) (also, even when there are 100 people visiting your website, the odds that all of them issue a request at exactly the same time is low - usually they load a page and read it - while they're reading it, some of the other people are loading other pages, and it all spreads out so you might only have ~10 page requests actively being processed by your server at the same time.)
However, if you have 2,000 people on your site at the same time, you might be processing 200 page requests at once, so by then you do need to have put some thought into performance and scalability.
(Note: these numbers are arbitrary and depend entirely on the type of site and it's traffic patterns.)
Generally, most websites pick a mid-level instance size, and then to handle more requests they 'scale out' - create lots of copies of that instance, and allow each instance to handle a portion of the traffic.
If I start another server say, Apache or Node, does that become another instance
The language to use here would be 'start another service say, Apache or Node' - they are other programs, and your instance will be perfectly fine running nginx, apache and node all at the same time. Although each will consume some of the resources (e.g. memory and cpu) and the more activity they are doing, the faster you will run out of resources and need to get a bigger instance size
So - no, they don't automatically become another instance. The language is confusing because sometimes people don't distinguish between the 'server' (aka the instance) and the service (aka the program) and will say the 'apache server' and the 'apache service' interchangably.
If I connect to a database, is that an instance?
Your instance, as a fully capable server, could run a database service on it at the same time as the other services - e.g. you could install and run mysql on your instance.
There is another option, though - if you use the AWS RDS product, then you will be starting an RDS instance. An RDS instance is different from an EC2 instance (what we've been talking about so far) in that RDS instances are specialised to just run the database service and nothing else, but EC2 instances are general servers that you can do pretty much anything on.
It's usually recommended to use RDS, but if you are trying to save money and aren't serving many users, there's nothing particularly wrong with installing mysql on your instance yourself (especially while you're learning how it works) and then moving your data to an RDS instance when you want to support more load or traffic.
Do instances start as needed?
Not by default, no - you have to manually start and stop them.
However, there are options other than manually starting and stopping. Amazon provides a lot of APIs, so you could write a program that would connect to the API and automatically start and stop your instance(s) based on rules you build into your program..
Also, Amazon offers a product called "AutoScalingGroups" which allows you to have a related group of instances and for Amazon to automatically start and stop them according to rules that you configure into that product. These rules can be 'scheduled actions' - start/stop at certain times of day - or they can be reactive - e.g. when the average CPU usage is > 50% for more than 5 minutes, start another instance.
This is more a side note, but I was reading this article and it said
For a client needing to run 800 virtual instances, the annual cost of
a private cloud came to below $400,000 vs. somewhere between $800,000
and $1.2 million for public cloud services.
The 'free tier' gives you a t2.micro sized instance (1 vCPU, 1 GiB RAM) which you could leave turned on permanently for free during that free year.
Even after your free tier expires, that same instance would cost you $9.67/mth, and you have the option to go downgrade to a t2.nano (0.5 GiB RAM) which would only cost ~$4/mth - but 0.5GiB RAM isn't much these days, so may not be enough for you.
A t2.micro should be more than enough to learn how to build websites on. If you are fortunate enough to build a site that is popular enough that you are getting more requests than that server can handle, then you will have to decide if you can generate revenue from that popularity sufficient to cover the cost, but by then you'll have more of a sense of how efficient your program is, and what instance size (and/or how many instances) you'll need.
Yes, it's obviously a big company, but can you imagine "hitting the
lottery" with an app everyone loves then before you know it, AWS hits
you with a bill of $1,000,000
AWS protects you from yourself here a bit - they have limits which generally restrict you from running more than 20 instances at a time - unless you ask for permission. So, by default, your instance won't go multiplying like rabbits on it's own - unless you set it up to. And even if you have set it up to, it won't be able to grow beyond 20 instances unless you have asked amazon to let you. So, worst case is 20 x $9.67/mth - $197/mth.
But - that's just the instance cost. Amazon charges you for lots of things including data traffic in and out, RDS instance costs, and if you start using other service such as S3 buckets and/or elastic load balancers, they all attract their own costs.
But hopefully, if you hit the lottery with an app everyone loves, you've worked out how to convert that love into dollars and cents so you can pay for all those instances you're going to need :)
This question has a conceptual and practical parts.
Conceptually I'd like to know if using the autoscaling functionality is equivalent to simply increasing the compute power by a factor of the number of added instances?
Practically ... how does this work? I have one running instance, its database sitting on an LVM composed of multiple EBS volumes, similarly with all website data. Judging from the load on the instance I either need to upgrade to a more powerful instance or introduce this autoscaling. Is it a copy of the running server? If so, how is the database (etc) kept consistent?
I've read through the AWS documentation, and still haven't got the picture yet - I could set one autoscaling group up which would probably clear my doubts, but I am very leery to do this with a production server.
Any nudges in the right direction would be welcome.
Normally if you have a solution that also uses a database, and several machines in the solution, the database is typically not on any of the machines but is instead hosted seperately with each worker machine pointing to the same database - if you are on AWS platform already, then DynamoDB or RDS are both good solutions for this.
In theory, for some applications, upgrading the size of the single machine will give you the same power as adding several smaller machines, but increasing the size of the single machine, while usually these easiest thing to do at first, should not be considered autoscaling and has its own drawbacks. Here are some things to consider:
Using multiple machines instead of one big one gives you some fault tolerance. One or more machines can go down and if your solution is properly designed new machines will spin up to replace them.
Increasing the size of a single machine solution means you are probably paying too much. If you size that single machine big enough to handle peak workloads, that means at other times (maybe most of the time), you are paying for a bigger machine than you need. If you setup your autoscaling solution properly more machines come on line in response to increasing demand, and then they terminate when that demand decreases - you only pay for the power you need when you need it.
When your solution is designed in this manner, you need to think of all of the worker machines as ephermal - likely to disappear at any time, so you need to build your solution differently. Besides using a hosted database (like on DynamoDB or AWS RDS), you also should not store any data on the machines in your auto-scaling group that doesn't also live somewhere else. For example, if part of your app allows users to upload images, you don't store them on the instances, you store them in S3. Same would apply to any other new data that comes in.
You need to be able to figuratively 'pull the plug' at any instant on any of the machines in your ASG without losing data.
Ultimately a properly setup auto-scaling solution will likely serve you better, but without doubt it is simpler to just 'buy a bigger machine' and the extra money you spend on running that bigger machine may be more than offset by the time and effort you don't have to spend re-architecting your solution to properly run in an autoscaling environment. The unique requirements of your solution will ultimately decide which approach is better.
I'm using cloud VPS instances to host very small private game servers. On Amazon EC2, I get good performance on their micro instance (1 vCPU [single hyperthread on a 2.5GHz Intel Xeon], 1GB memory).
I want to use Google Compute Engine though, because I'm more comfortable with their UX and billing. I'm testing out their small instance (1 vCPU [single hyperthread on a 2.6GHz Intel Xeon], 1.7GB memory).
The issue is that even when I configure near-identical instances with the same game using the same settings, the AWS EC2 instances perform much better than the GCE ones. To give you an idea, while the game isn't Minecraft I'll use that as an example. On the AWS EC2 instances, succeeding world chunks would load perfectly fine as players approach the edge of a chunk. On the GCE instances, even on more powerful machine types, chunks fail to load after players travel a certain distance; and they must disconnect from and re-login to the server to continue playing.
I can provide more information if necessary, but I'm not sure what is relevant. Any advice would be appreciated.
Diagnostic protocols to evaluate this scenario may be more complex than you want to deal with. My first thought is that this shared core machine type might have some limitations in consistency. Here are a couple of strategies:
1) Try backing into the smaller instance. Since you only pay for 10 minutes, you could see if the performance is better on higher level machines. If you have consistent performance problems no matter what the size of the box, then I'm guessing it's something to do with the nature of your application and the nature of their virtualization technology.
2) Try measuring the consistency of the performance. I get that it is unacceptable, but is it unacceptable based on how long it's been running? The nature of the workload? Time of day? If the performance is sometimes good, but sometimes bad, then it's probably once again related to the type of your work load and their virtualization strategy.
Something Amazon is famous for is consistency. They work very had to manage the consistency of the performance. it shouldn't spike up or down.
My best guess here without all the details is you are using a very small disk. GCE throttles disk performance based on the size. You have two options ... attach a larger disk or use PD-SSD.
See here for details on GCE Disk Performance - https://cloud.google.com/compute/docs/disks
Please post back if this helps.
Anthony F. Voellm (aka Tony the #p3rfguy)
Google Cloud Performance Team