SQL Server 2016 is vastly different from its predecessor SQL Server 2014, with a set of cloud-first capabilities coming to on-premise and designed to enable future-ready enterprises.
“It is not just another database that we are launching. The birth of this product took place in the cloud, so whatever features we have added in SQL Server 2016, have already been tested and used by the customers in the last one and a half years on Azure,” says Srikanth Karnakota, Director – Server and Cloud Business, Microsoft India. SQL data warehouse capabilities and others which have been brought in to SQL Server 2016 on-premise is something Azure has already been working in. Thus, it’s cloud first capabilities are coming back to on-premise.
The platform different and more precise than the one launched in 2014. When it comes to security, with SQL Server 2016, data is always encrypted, whether data is in motion. For example, when doing a transaction with a credit card, once a customers click and feeds data, there is a transit time before data gets stored into a database and then there is a validation happening. At that time, data is not encrypted and one is still vulnerable. Once the data hits the database and there is a hack on to the database, then also the data is vulnerable. “What SQL Server 2016 enables is, whether the data is in transit or written onto the database or the data is being retrieved later by some other app; at every level it is encrypted. The public key of that resides with the customer and if someone does hack into the database server without a public key they cannot make out any sense of the data. Customers can store public key wherever they want and even a database administrator would not be able to find the data that is inside,” Karnakota explains. This is the technology that Microsoft Research has been working for the last 4 to 5 years.
Since this database took birth in cloud it has been battle tested and lends itself to cloud extensibility seamlessly. It means that it has features like Stretch. Its Stretch capability has been touted as a game-changer and Karnakota digs deeper justifying why it is so? All large enterprises have major set of data of which 80-90% is not used meaningfully. But they end up storing their data on expensive SAN systems. The ideal place to store data which is not accessed often is cloud, because that’s the cheapest way to store any data. Imagine having the power in the database itself that is intelligent enough to figure it out that you don’t use that data so often and the smartest place to store such a data is to ship it to public cloud which is Azure. The problem if you store the terabytes of data on public cloud is query. SQL Server 2016 allows customers to query the database sitting on the cloud and there is no compromise on quality, performance and security. No other database can talk this language. ”
Exploring the Stretch capability further he gives an example saying, “A retail company who wants to have a database of all the customers who walk in, might or might not have a loyalty card with them. The loyalty card data would be residing in the IT systems. So whenever this customer walks in to purchase, the retailer can see what do they buy or spend, how many times and all this data comes with it. Now, if you want to analyze that what is the buying behavior pattern of customers who walk in to the store one can superimpose all the consumer data around that retail store on the already available data. This could be the data which is not accessed very often as well.
The developer building an app on Stretch will literally be able to say which data is not accessed often and have a simple trigger in the data setups to move this data to the cloud and when it needs to be analyzed, make it readily available. So the way people write application changes, when they start using Stretch.
He further expounds on it informing, “But what is more powerful than that is the cloud platform when it comes in the picture. For example, the CIO doesn’t have any data on what customer does outside because he doesn’t know how to collect. But there are other data sets which are outside the enterprise like Facebook, Twitter, Consumer Survey of India which analyses the buying patterns.
That data put on a public cloud has algorithms which can be combined with structured data that is sitting within the enterprise which is coming through loyalty cards and then juxtapose both of them. And when you juxtapose both of them these are the additional insights that the retailer gets. Now what does a retail company do after that? He has a beautiful program and he is able to target better. That’s precise information.”
Contemplating on his favorite feature in SQL Server 2016 Karnakota talks about performance. He says, “Column Store Indexing is able to drive almost 30 x performance improvement in how transactions get processed in a database. So what happens with other database companies is they would insist you to get very expensive hardware to achieve the kind of performance output. But imagine, if the software has the capability to do parallel processing and the way one stores the data; one can index and query it so fast that commodity hardware is able to give you 30x performance improvement.
Large banks want to do churn analysis on all the users, they want to say that which of the customer is not going to be with the bank next month. They need to be able to do that especially in region where there are high churn ratio. You will never run that query on a system that is already interacting day and out with customers, because performance deteriorates for the customer when he tries to access it. So what ends up happening is the mission critical applications that are running that is one part of your enterprise, the analytics part is another enterprise, that is why they take data from cubes and do data warehousing in a separate system. The reason it’s happening is there is no performance ability in the system to be able to serve customers and at that same time be able to do analytics on its own. That’s the capability that SQL Server 2016 brings with Column Store Index.
When it comes to Hybrid cloud, it is a highly differentiator platform. The rich data visualization combined with the power of cloud and database which resides locally completes the entire data platform. “Seven companies need to come together to have such capabilities and these are the capabilities I am most proud of,” he points out thoughtfully.
There are definitely some tips and recommendations that customers can leverage before embarking on their SQL Server 2016 journey. Karnakota gives recommendations on three levels. First is to be very clear about what workloads an enterprise is going to run where. The second recommendation is that if an enterprise already has a platform which is on-premise and want to connect it to the cloud, they can look at SQL Server 2016. Microsoft has made an offer for all customers who are running on Oracle, of getting free SQL Server 2016 in terms of licensing. They just need to pay for maintenance.
The third recommendation in terms of how customers embrace it is workloads and having the right set of partners and platform companies to work with. Enterprises should look at which workloads they want to prioritize like customer churn, predictive intelligence, being able to tell how your next quarter is going to do etc. Also, which applications do they want to modernize, which applications do they want to build grounds up and which can lend themselves to cloud?
Amidst applauds there are also mixed reactions to SQL Server 2016 running on Linux which is said to have only certain capabilities and not all capabilities. Explaining it Karnakota says, “Right now we are looking at what all versions of Linux we are going to support. The feature set of SQL that will run on Linux and Windows will predominantly be the same. But there are certain versions of Linux where you will have these gaps and we are going to certify which versions of Linux we are going to come up with.”
Speaking on SQL Server 2016’s direct business implications Karnakota summarizes, “The new world in a Telco is looking like what is the propensity for a customer who buys a prepaid card and when does he do a top up and how much will he top up. There has to be a new application to do that. The largest bank is worried about what is happening with e- wallets that are coming in, because they will disrupt the banking industry. The next generation of Pharma company is not about that ERP, Supply-chain etc. The next gen of Pharma company is about which markets to innovate with new drug discovery. So by industry we are realizing these two world coming together… the cloud world and super analytics in the data world community. The whole data explosion that is happening is creating a whole category of applications that are super mission critical for the future and that is the world for which SQL Server 2016 is designed for.”
