Skip to main content

Scaling Units - Having resources outside instances of Scaling Units

This year I had different posts where I talked about Scaling Units and different approaches of implementing Scaling Units. In this post we will discuss why it is important to not have shared resources across different scaling units (like storage, database or messages systems).

Before going further let's see why we are using Scaling Units.
We are using Scaling Units to be able to ensure the same quality attributes of a system for 10, 10.000 or 1.000.000 users. This is the most important thing that Scaling Unit is offering to us.
More about scaling units can be found here: http://vunvulearadu.blogspot.com/2015/02/scale-units-and-cloud.html

A Scaling Unit contains resources that are dedicated to a fix number of users or devices. If the maximum load is exceeded, a new Scaling Unit is added. Internally, in a Scaling Unit we will never add more resources to be able to manage a higher number of users or devices.
Of course, when we implement such a system, we will identify resources that are shared or can be shared between them.
For example, the repository where binary storage is stored could be shared between them Having for example a repository of pictures that needs to be accessible by all of our users across all region can make we think that this is the perfect resources that can be shared across Scaling Units.
The question that is coming now is: "Should we shared this type of resources between Scaling Units?"
There is no perfect answer, like when we implement OOP, there is no perfect solution. Based on our needs, costs and requirements we need to find the best solution.
The base idea behind Scaling Units is to not share any kind of resources between Scaling Units. By doing this we will be able to ensure the best quality of our system and to ensure the same quality attributes. Of course this solution would come with a cost and there are times when we don't need such a solution.
In the above example there is no resource shared between them. We can ensure the minimal download bandwidth for pictures because we know the number of users that are managed by each Scaling Unit. By knowing all the parameters, we can measure the load, simulate it and reserved the necessary bandwidth that is required to offer the required quality attribute.

When the quality attributes are more relaxed, or because of the context we can share resource between Scaling Units. Even if this is not recommended and I'm not a fun of this approach - the world that we leaving in is not an utopian world.
There are different approached to share resources. I will present two of them, that I saw on different solution and a 3rd one that I implemented in the last project. The last one is the one that I think that is the best way of doing it.
The first solution is the most simple one. We take the resources from a Scaling Unit and we shared with the other instance of Scaling Unit. The resources will be shared between two or more instances.
All the time we know which Scaling Units has access to that resources.
The second approach is when the resources is shared between all the Scaling Units. In that moment the resources is part of all Scaling Units and we can only hope that quality attributes that we need to offer will remain in the green parameters.


The 3rd way of sharing resources between Scaling Units is to extract the resources outside the Scaling Unit. That resource will not be part from any Scaling Unit. Will be a separate resource, that is managed and scale separately. Very similar with a micro-service. There can be one or more instances of that resource. The access and allocation at this resource is controlled by an external system (that is not part of our Scaling Units).
The instances of our resource should be allocated to an instance of a Scaling Unit not to specific users of our Scaling Units. We can have for example an instance of our Scaling Unit that is allocated to 2 instances of our Scaling Units. The Scaling Units will know exactly at what instance of our resource should go.

When there are issues with that resource, Scaling Units should go to the controller of our resource to request access to an instance of the resource.
By extracting the resource outside the Scaling Unit we have full control of that resource, we can scale it separate, based on the needs. We can ensure the same quality level across different scaling units.

Conclusion
Extracting resources outside Scaling Units is not recommended, because we are losing the benefits of Scaling Units. When this is done, it needs to be done with precaution and by respecting the same concepts as for Scaling Unit. Otherwise we will not have extra value from using Scaling Units.


Comments

Popular posts from this blog

Windows Docker Containers can make WIN32 API calls, use COM and ASP.NET WebForms

After the last post , I received two interesting questions related to Docker and Windows. People were interested if we do Win32 API calls from a Docker container and if there is support for COM. WIN32 Support To test calls to WIN32 API, let’s try to populate SYSTEM_INFO class. [StructLayout(LayoutKind.Sequential)] public struct SYSTEM_INFO { public uint dwOemId; public uint dwPageSize; public uint lpMinimumApplicationAddress; public uint lpMaximumApplicationAddress; public uint dwActiveProcessorMask; public uint dwNumberOfProcessors; public uint dwProcessorType; public uint dwAllocationGranularity; public uint dwProcessorLevel; public uint dwProcessorRevision; } ... [DllImport("kernel32")] static extern void GetSystemInfo(ref SYSTEM_INFO pSI); ... SYSTEM_INFO pSI = new SYSTEM_INFO(

Azure AD and AWS Cognito side-by-side

In the last few weeks, I was involved in multiple opportunities on Microsoft Azure and Amazon, where we had to analyse AWS Cognito, Azure AD and other solutions that are available on the market. I decided to consolidate in one post all features and differences that I identified for both of them that we should need to take into account. Take into account that Azure AD is an identity and access management services well integrated with Microsoft stack. In comparison, AWS Cognito is just a user sign-up, sign-in and access control and nothing more. The focus is not on the main features, is more on small things that can make a difference when you want to decide where we want to store and manage our users.  This information might be useful in the future when we need to decide where we want to keep and manage our users.  Feature Azure AD (B2C, B2C) AWS Cognito Access token lifetime Default 1h – the value is configurable 1h – cannot be modified

What to do when you hit the throughput limits of Azure Storage (Blobs)

In this post we will talk about how we can detect when we hit a throughput limit of Azure Storage and what we can do in that moment. Context If we take a look on Scalability Targets of Azure Storage ( https://azure.microsoft.com/en-us/documentation/articles/storage-scalability-targets/ ) we will observe that the limits are prety high. But, based on our business logic we can end up at this limits. If you create a system that is hitted by a high number of device, you can hit easily the total number of requests rate that can be done on a Storage Account. This limits on Azure is 20.000 IOPS (entities or messages per second) where (and this is very important) the size of the request is 1KB. Normally, if you make a load tests where 20.000 clients will hit different blobs storages from the same Azure Storage Account, this limits can be reached. How we can detect this problem? From client, we can detect that this limits was reached based on the HTTP error code that is returned by HTTP