We are a solution provider and SQL Server is one of the products that we implement for our clients. People use SQL Server primarily for business intelligence.
If you have Navision or Dynamics AX, and you need a data warehouse to be able to improve the performance of your business, then that is what our company does.
For this review I will describe one of our clients. They are a large company that owns convenience stores, and they do several billion euros a year in business world wide. The convenience stores are located in large travel hubs like airports, train stations, office buildings, and large supermarkets. Everyone who flies/travels a lot has shopped in their stores. In Romania, they have 300 stores and sell between 300,000 and 500,000 items a day across the network.
This customer is very mature and we have provided a lot of features. The most important is forecasting. When the store manager comes to the store in the morning they receive an order proposal dashboard from us. We will have taken the data from the close of business the night before, then built a forecast for the manager to help decide what to order the next day. This a very important application. We are able to calculate the orders across the network better than the ERP can.
Another use case has to do with the actual items in convenience stores. They sell products such as cigarettes, drinks, sandwiches, magazines, and more. In Eastern Europe, there's still a high percentage of people that smoke cigarettes, and what you want to do, as a retailer, is negotiate your contracts with the large vendors.
There are people called buyers, and they sign contracts with the major suppliers. Our use case provides a tablet with a very rich dashboard to show everything that is going on between the buyer and the vendors that they buy from. In those negotiations, especially when dealing with very larger vendors, you sell millions of euros of goods, such as cigarettes, each year. The goal is that you are trying to get the vendor to give you a lower price. At the same time, the vendor is trying to get you to pay a higher price. That's business.
In that discussion, the person with the best data gets the best price. Naturally, it's a high-conflict discussion and you have to have a way of taking out the conflict and replacing it with a more logical, rational argument. The person with the best data wins the argument and gets the best price. That capability is worth a lot of money. Although only a dozen people use it, it's one of the top applications used by that customer.
One of the most valuable features is partitioning because it helps to distribute the workload. Consider that you sell 400,000 or 500,000 items each day, and there are 10 years worth of history stored in the data warehouse. That is a lot of transactions. Partitioning allows you to split it up into smaller pieces so that the machine can be easily deal with it.
By default, SQL Server has relatively poor partitioning, and it only works properly if you buy the Enterprise Edition, which costs about $14,000 USD per core. Partitioning was only introduced in 2005 and because we were working with it from before that time, we had to do partitioning manually. With Standard Edition, the list price is $3,700 per core. So we still do partitioning manually.
With SQL Server SE, for a $50,000 USD or so license, we can get a 16 core server and manage data warehouses in the 15-20TB range. It runs really well.
By comparison, we had a large telecom client in 2004. We used an 18CPU Sun Server running Oracle and 15TB of EMC disk. That cost us nearly $2.5 million USD. Nowadays, we can do a great deal of what we did back then with $100,000 USD. This is amazing because in places like Eastern Europe where you have constrained budgets SE is a really good option. It's actually cheaper than running MySQL.
