Azure Managed Instance

First, The Early Days!

The early days of SQL Server Azure PaaS offerings were a period of change for data professionals, DBAs, etc. As a lot of colleagues and end users remarked “Where is the SQL Agent? Life simply isn’t the same anymore!”

Well truly, life as a SQL Server data professional was not the same anymore. Azure SQL Server PaaS brought about a paradigm shift in how we perceived SQL Server database options and provided malleable offerings for various scales of implementation. Azure SQL database, Elastic pools became conversation starters for database implementations or migrations. Then came March 7, 2018 when the customer preview for SQL Database Managed Instance was announced and we saw the words, “Almost 100% feature compatibility with full SQL Server installations” and yup, that’s right, we had the SQL Agent on PaaS . A form factor of SQL Server that we have been familiar with for decades is now a PaaS offering.

Managed Instance General Purpose tier general availability was October 2018 and the Business Critical tier general availability was December 2018 ensuring a range of offerings catered to various sizes of deployments as well as various performance and availability requirements.

Why Managed Instance?

Managed Instance is a fully kitted out SQL Server PaaS option which offers up almost all the feature set of a full SQL Server database installation. This makes application database feature parity easier, and also eases up the way for “Lift and Shift” database migrations from existing SQL Server database installations. All these advantages along with built-in Azure features of a fully managed service, platform ingrained high availability, performance monitoring, secure network implementations made Managed Instance a truly worthwhile alternative to SQL Server on Azure IaaS or the good old on-premise SQL Server installations.

What makes it tick?

A comprehensive feature set that includes service broker, native backup and restore, Query Store, transactional replication, change data capture etc.
As alluded to earlier, the ease of moving an application database with little code change ensuring compatibility with current on-premise offerings.
A scalable implementation that allows scaling up of vCPU (up to 80 vCPU) and Storage (Up to 8 TB for General Purpose tier, 4 TB for Business Critical tier) allowing quicker resource provisioning turnaround times than an on-premise set up.
An out of the box fully managed high availability solution assuring 99.99% uptime.
Existence of a Business Critical tier catering to low I/O latency requirements even for large storage requirements.
Managed Instance auditing allowing an inbuilt method to ensure security and audit compliance that plays well in a stricter regulatory compliance climate in the market.
Data encryption in motion and Dynamic Data Masking through policy-based security implementation.
Did we talk about Vnet? Yup, we need to. The only Azure database PaaS option offering network isolated database implementations allowing secure isolated database instances accessible through VPN or Express route.

Managed Instance, show me the money!

Well, the move to PaaS should ultimately realise cost savings in addition to the operational efficiencies the platform delivers inherently. The savings can be attributed to a few factors:

Azure Hybrid benefit

Like most Azure SQL database offerings Managed Instance benefits from cost saving brought forth by Azure Hybrid benefit bringing about an approximate 30% savings in licensing costs. Interestingly Microsoft announced Azure Hybrid benefit at the same time as the Managed Instance general availability.

Operational cost savings

With no longer the need to provision, maintain and upgrade the compute hardware and storage on-premise and having offloaded these to Microsoft, organisations have significant opex savings realised. Considering the size of typical Managed Instance workloads, the savings are magnified by quite a few multiples.

Improved DBA productivity

The move to Managed Instance freed up significant management over from the DBAs and allowed organisations to deploy their DBAs towards driving efficiencies through automation and new technology initiatives. With the trend of organisations moving down the DevOps path becoming the norm, freeing experienced and skilled DBAs from mundane tasks and focusing on driving efficiencies is a win-win all around.

Database lifecycle management costs

With no need to plan, schedule or execute database upgrades that are now taken care of by Microsoft by getting the latest SQL Server builds and versions available on Azure Managed Instance. This leads to significant technology budget outlay savings on upgrades.