Skip to main content

Azure VMs with SSD are now available - Benchmark using MongoDB over SSD

Today, Azure team announced that we will have VMs with Solid State Drives. Wow, this is so cool.
Why this is so great? Well, first of all, if you work with application that have a high demand of I/O on disk than this could be the perfect solution for you. Of course, the new machines has a new configuration, with additional processing power, but I don’t want to talk about this.
For more information related to this VMs you can go to:


In this blog post we will focus on the new SSD that are coming on this machines. Based on the configuration, the new D-Series machines are coming with local SSD that start from 50GB and goes to 800GB.
It is pretty interesting to put a MongoDB server on this partitions or other intensive I/O application. Also, if you have SQL Server on a VMs, you can put BPE (Buffer Pool Extension) on that partition. Don’t forget that BPE can be activated very easy and can increase the performances of SQL Queries. I expect a performance improvement of this kind of applications.
An important thing that you need to keep in mind. This SSD drives are only available for the TEMP drive. The operating system will still run on a classic drive. In future we may see other parts of VMs running on SSD, but is only a dream.
Additional to this, because we are in the TEMP driver (and we are using a SSD), you should add only data that you can afford to loss.
Benchmarks
I found an interesting benchmark related to Microsoft Azure VMs that I think that could be interesting for us:

  • A1 VM (1 core, 1.75GB RAM) Score: 2703
  • D1 VM (1 core, 3.5GB RAM) Score: 3947
  • A4 VM (8 cores, 14GB RAM) Score: 8705
  • A8 VM (8 cores, 56GB RAM) Score: 17186
  • D4 VM (8 cores, 28GB RAM) Score: 12645
  • A9 VM (16 cores, 112GB RAM) Score: 29873
  • D14 VM (8 cores, 112GB RAM) Score: 22017 

MongoDB Load Test
Because in the last period of time I had the opportunity to play a lot with MongoDB, I had the setup of a load test for MongoDB. Nothing special, around 30M objects in one table. I only execute read operations over MongoDB, without any writes over it.
I run this load test 20 times on this machine, each time making a full restart of the machine. 10 times I had the database in the temp folder (SSD support) and 10 times in a normal folder (standard hard disk).
The results were pretty interesting, from my perspective. SSD disk improved with around 2.5X the performance of MongoDB queries. The result time is not very relevant, but the average execution time reduced from ~30ms per query to around ~13ms per query.
Take into account that SSD are not so fast when we need to make an insert or update. Because of this I didn’t even wanted to do a test like that. I don’t like the ideas to store content in a directory where I can loss all the information (like TEMP), but using a replica you can resolve this issues (but the performance will be lower).
Conclusion
There are scenarios when a SSD can help you a lot when you need to improve your performance, especially when you need to make a lot of I/O operations (especially read). I like the idea to have VMs with this kind of configuration and I expect to use them in the near future.

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