Skip to main content

What are the limitation of Windows Azure Tables

I’m pretty sure that a lot of you had heard about Windows Azure Tables. I already described how we can work with this Windows Azure Tables. You can find a series of posts about them on my blog in this link.
Today I want to talk about some limitation that we can have on Windows Azure Tables if we don’t use properly. A row in a table can store any kind of data that is serializable and one table can have more than one entity type saved. For example in the same table we can save Student entities and in the same time Dog entity and also Car entity. Each entity that is saved has 3 properties that need to exist all the time:
Partition key – based on this property we can group items from a specific table based on this partition. It is used for load balancing across storage nodes.
Row key – this a unique identifiers used to identify a unique row in a partition (the combination between partition key and row key form a unique key in the table).
Timestamp – the last modified time for a row.
The partition key exists with a scope, but a lot of people don’t use it properly. Because of this when they do intensive work on a table from Windows Azure some performance issues could appear.
Why? In this moment (July 2012), it seems that the maxim number of rows that can be processes per partition in a table is 500.
What does it means? If we have a table with only one partition key we will be able to process only 500 rows per second. If we have a table with 10 partition keys and we will be able to access maxim 5000 rows per second.
Some people could say: “OH, only 500 rows per second!”. What we don’t need to forget is that all this information’s are usually access from internet, not only from the datacenter. Because of this sending for example 500 rows per second is a lot of information and even on our side (on the client that consume the table) we could have problems with our bandwidth.
Also, we can use partition key in such a way that we could distribute the rows in a table in a manner that this limitation will not be reached.
For example if we have a table where we store audit data, we could group them based on the type. For example each partition key would represent a different audit type. A better solution is to have each audit time in a different table and the partition key could be used to group the audit data based on the type (per hour, per day, depends on how many items are logged per hour/day). Also, don’t forget that we will pay the same amount of money if we have one Azure table with 1.000.000 or 100 Azure tables with 10.000 rows each.
Don’t be afraid to split information in more than one tables. Having data in more than one table will help us to work with it more easily. Also use partition key, whenever is possible. We have them with a scope.
In conclusion, we don’t need to forget that we have a limited number of rows that we can process per second. Now are 500 rows per partition key, in the future maybe will be 10.000. What is important for us is to know that some limit exist and when we use table intensive, we need to remember that we need to distribute data across table or partitions.

Comments

  1. Interesting - what is important to remember is that Azure table storage is a NoSQL DB - one good way to decide which entities should be grouped in the same partition is to think in terms of aggregates (from DDD) - if the partition key is the same for all entities in the same aggregate, the 500 rows/s limitation is more than enough for most usual scenarios.

    Another important purpose of partition key is that it's used to make atomic insert/update possible in Azure - all the data in the same partition is stored physically close.

    ReplyDelete

Post a Comment

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