The market for data warehouse platforms looks set to expand as gleaning insights from data becomes ever more important for businesses. At the same time, established names like Teradata are struggling lately against Amazon and Microsoft and some well-funded newcomers, especially those offering the convenience of a cloud service.
The overall size of the data warehouse market depends upon who you ask, but a recent report published by Allied Market Research pegged it at $18.61bn in 2017 and estimates it will be worth $34.69bn by 2025.
However, the spoils will not be shared equally. Let’s take the example of one big industry Teradata, which has seen revenue fall over the past few years, down from around $2.7bn five years ago to $2.2bn in 2017, as reported in The Register.
The decline can be attributed to several factors, such as customers wanting analytics as well as data warehousing, analytics in the cloud and pay-as-you-go cloud-style pricing models.
Meanwhile, the Hadoop vendors who rode the big data analytics wave at about the same time are also struggling. Two of the biggest players, Cloudera and Hortonworks, announced a merger last year after both recorded losses.
The post-merger Cloudera has also just seen its share price fall as it missed revenue expectations and new CEO Tom Reilly quit after a few months in the post.
Snowflake Computing and Yellowbrick Data seem to be doing better. These privately held startups do not reveal sales figures, but backers are piling megabucks into both companies. Snowflake has raised $923m to date, while YellowBrick Data this week secured $81m in Series C funding that brings its total up to $173m.
It’s the cloud wot done it
If you guessed that cloud had much do with these changes in the data warehouse market, then you would be right.
“It is indeed the cloud. The market as a whole is very aggressively starting to pivot towards the cloud,” said Adam Ronthal, senior director analyst for data management and analytics at Gartner.
He noted that just two vendors – AWS with Redshift and Microsoft with Azure SQL Data Warehouse – accounted for 70 per cent of the revenue growth in the overall DBMS market for 2017 and 2018 .
The usual reasons apply – it is easier to engage with cloud-based technologies, so the barriers to entry are significantly lower. Customers do not have to install or maintain complex and costly infrastructure, and billing is more transparent.
Riding this wave are upstarts like Snowflake that pitch their platform as cloud-native without any legacy baggage.
Traditional vendors have responded to this challenge by either repositioning their platforms, or by trying to adapt them to run in the cloud.
Teradata is a good example, and is in the midst of repositioning itself at the very high end of the market, where people have the most challenging and complex workloads, according to Ronthal.
This effectively means the company is retreating to a market niche, and the Redshifts and Snowflakes have a good enough proposition for the majority of customers.
Yellowbrick Data is currently an on-premises play, but it came into the market promising higher performance for customers with legacy platforms such as Netezza, and in some cases claims to have replaced multiple racks of older equipment with a single Yellowbrick enclosure.
The company now touts itself as a hybrid cloud play by aiming to get its platform deployed on public clouds, so that Ronthal says that he regards both Yellowbrick and Snowflake as riding the cloud wave.
As for the Hadoop companies, they are suffering a hangover from over-hyping five to six years ago. At the time there were predictions-a-plenty that non-relational databases would become a universal platform to replace everything else. Instead, Hadoop and NoSQL are finding their place alongside the traditional SQL-based platforms.
According to Ronthal, “there has been a rediscovery of the core value of some of the relational technologies which are still best suited for production delivery and high degrees of optimisation, whereas the non-relational technologies have their place but they are not the same value proposition,”
The overall picture is that data warehousing is shifting inexorably to the cloud, but customers still look for platforms that offer the traditional relational database tools at their core. This is good news for Snowflake and Yellowbrick.