Teradata provides the lowest cost per query for enterprise-scale analytics. I know that many of you may find that surprising, so read on…
I find it fascinating that many people discuss the analytics market as if it is one segment. It would be like trying to discuss baseball without knowing if it is major league, minor league, or little league. You have to dig deeper…
The latest hype cycle is all around “less expensive” cloud-native data warehouses and pay-as-you go pricing. These “cloud-only” software vendors tout their solution as a panacea for all customer requirements. In reality, there are two macro-segments that are vastly different. One size does not fit all, and this includes software functionality, deployment model, and pricing model.
Macro-Segmentation Model
The first segment is Basic Data Warehouse. It is usually application-specific and serves a focused group of users, such as a department. Almost any new system in this segment will be in the cloud. Most of these workloads are sporadic and can be turned off when not being used. The user concurrency is relatively low, in the tens to hundreds of queries per day. The cloud-only software vendors have made it easy to scale up the underlying compute infrastructure, sometimes automatically. And these systems are painfully simple to operate – no tuning required. Their analysis is based primarily on SQL-only for reporting and business intelligence. Customers in this segment only want to “pay for what they use.” Since basic data warehouses are fairly simple, the business value generated is low to moderate.
The second segment is Enterprise Analytics. Customers in this segment architect for an enterprise scope by integrating cross-functional data that serves multiple applications and user groups. The majority of these systems are on-premises today since most source data comes from on-premises operational systems. These systems are not intended to turn off and back on. In fact, they cannot go down since they serve operational users and SLAs 24 x 7. The concurrency can be extreme in this segment, sometimes running over 50 million queries per day, hence customers cannot afford to pay per query. Instead, customers require a stable and predictable pricing model. Robust workload management capabilities are necessary, and these customers want to exploit advanced data science functions in addition to SQL. These environments require performance tuning via advanced indexing and sophisticated query optimizer techniques, in addition to hardware scaling. The business value generated by these enterprise analytics systems is extremely high.
You will also find that many companies start with basic data warehousing and naturally evolve (if they are successful) to more of an enterprise analytics approach over time. This is where things start to break down for some companies that start with technology that is not flexible. As more users and applications are added, you can quickly see cost surprises. And if the software is not capable of meeting advanced requirements, you may be left with having to change the platform, which is extremely disruptive and limits the business value you can realize.
.png?origin=fd "Picture1-(1).png")
Customers are choosing Teradata because of the flexibility of Teradata’s Vantage platform. We allow customers to seamlessly move from a basic data warehouse to the most advanced enterprise analytics with no disruption. Vantage runs in multi-cloud and hybrid environments. Customers are not forced to only run in public clouds. They run their analytics wherever they need to and can easily move across deployment options.
Teradata Vantage also offers choice of pricing models. Customers can start with no up-front commitment and only pay per query. As they scale up, they can switch to a more traditional capacity pricing model. And Vantage has the power to go far beyond basic SQL since we’ve integrated sophisticated data science and advanced analytical functions natively into the platform
In a recent study, Price Performance in Modern Cloud Database Management Systems by McKnight Consulting Group, the premier benchmark experts for Cloud Analytics, some critical insights emerged:
Data Warehouses may look great at the low end during vendor-crafted POCs, but in reality implementations can get very expensive if deployed at production scale. Some organizations have seen upwards of 3-5X higher costs than originally projected. This has forced some deployments to different solutions – either on-premises or a different cloud platform. Time wasted. Budgets impacted. Migrations disrupted. It doesn’t have to be this way.
While no one wants an over-engineered system, there is only upside to having indexes and tuning at your disposal. Free performance features are the sign of a mature database. They enable service level commitments and workloads that cannot not be achieved otherwise. No tuning is often great for marketing, but impractical in the real world and not in the company’s best interest.
We are starting to see customers who purchased a cloud-native basic data warehouse disillusioned with their actual costs as they attempt to stretch it into an enterprise analytics system. What started off as a simple, cost-effective, pay-as-you-go model gets hard to tame when the usage goes up. Their costs are escalating each time their hardware auto-scales. Some of these were Teradata customers who were promised cost savings to switch and they are now re-committing to Teradata. This is why Teradata provides the lowest cost per query for enterprise-scale analytics.
