Monday, September 30, 2019

Azure Dedicated

I remember a discussion with a group of friends around 8 years back. Microsoft was in it’s early days of becoming a leader in the cloud. Those friends, all techies in the Seattle area had varying expectation on how it would work out. Many thought that a full blown move was few decades away because their experience indicated that all big companies ran on stack that was very old and simply couldn’t be moved to the cloud any time soon.

Waves of workloads have been since moving to the cloud. A new brew of startups were cloud native from the start and they were the first to use the power of cloud. Many large and small enterprises had already virtualized workloads and they moved as well. Some moved their new workloads (green-field), some even followed lift-n-shift with some modifications (brown-field) into the cloud.

However, a class of large enterprises were stuck in their data centers. They wanted to use the power of the cloud, they wanted to use IoT integration, Machine-learning and the capability of elastic growth of their applications, but the center of their systems were running on some stack that did not run in the standard virtualization offered in the cloud. These enterprises said that if they cannot move those workloads into the cloud, they would need to keep the lights on in their data-centers and moving some peripheral workloads simply did not make sense.

This is where Azure Dedicated and we come into the picture.

SAP HANA Large Instance

For some of these customers that #$%#@ is SAP HANA in-memory DB on a single machine with 768 vCPUs and 24 terabytes of ram (yup) and we have them covered. Some wanted to scale those out to 60 terabytes in memory, we have them covered too with our bare-metal machines running in Azure. See SAP HANA Large Instances on Azure. They wanted to then expand their applications elastically using VMs running on Azure with sub 1 ms latency to those baremetal DB machines, we have that working too.

We started our journey in this area with this workload. Now we have evolved into our own little organization in Azure called Azure Dedicated and also support the following workloads.

Azure VMware Solutions

Some customers wanted to run their VMware workloads and we have two offers for them, see more about Azure VMware Solution by CloudSimple and Virtustream here

Hardware Security Modules

In partnership with other teams in Azure we support HSM, which are standard cryptographic appliances powering say financial institutions.

Cray Supercomputer?

So you need to simulate something or do ML on tens of thousands of cores, we have Cray super computers running inside Azure for that!!

Azure NetApp Files

Working closely with the storage team we deliver demanding file based workloads running on Azure NetApp files

What next?

We know there are more such anchors holding back enterprises from moving into the cloud. If you have some ideas on what we should take on next, please let me know in the comments!