Home / News / SQream Technologies raises $39.4 million for GPU-accelerated databases

SQream Technologies raises $39.4 million for GPU-accelerated databases

Knowledge set analytics startup SQream Applied sciences these days introduced it has raised $39.four million. The majority of the brand new price range will cross towards ability recruitment, product R&D, and adorning the corporate’s buyer supply platforms, in line with a spokesperson. SQream targets to decrease the barrier to important knowledge optimization with higher efficiency, a discounted footprint, and value financial savings over its competition.

Forbes discovered that simply over 50% of businesses are adopting large knowledge analytics, with 95% bringing up the want to set up unstructured knowledge. This uptick could also be most likely resulting from ROI adopters file seeing — Entrepreneur notes that companies the use of large knowledge enjoy a benefit build up of eight% to 10%, on moderate. However hindrances stay, with 40.three% of respondents to a Statista survey suggesting large knowledge utilization was once held up by way of a loss of organizational agility.

SQream’s column-oriented database product — SQream DB — is a GPU-accelerated knowledge warehouse in a position to dealing with complicated queries with options cherry-picked from relational database methods. The usage of on-premise and cloud acceleration, SQream claims its shoppers analyze trillions of information and cargo as much as three terabytes in step with hour in step with GPU out-of-the-box.

SQream DB decouples garage from compute, getting rid of the requirement to duplicate, redistribute, or repartition knowledge when groups develop. It integrates with industry intelligence equipment by the use of a wide array of drivers and connectors, together with ODBC, JDBC, Python, Node.JS, Spark, R, Java, and C++. SQream DB’s interface layer serves as a choice of products and services that keep an eye on the knowledge warehouse, whilst the compute layer is the place the true knowledge processing duties are run. The closing layer — the garage layer — is divided right into a metadata module the place all regimen database gadgets are saved and a power bulk knowledge element that’s closely optimized for uncooked table-scan efficiency.

VB Change into 2020 On-line – July 15-17. Sign up for main AI executives: Sign in for the unfastened livestream.

SQream Technologies

SQream DB can learn knowledge immediately from exterior resources with the exterior desk syntax, and its columnar garage machine is partitioned horizontally and vertically for analytic operations like joins, aggregations, summarizations, and sorting. The columnar engine lets in selective get right of entry to to the specified subset of columns, decreasing disk scan and reminiscence I/O as opposed to same old garage. In the meantime, hyper-partitioning splits up garage horizontally into manageable chunks, complementing AI-assisted compression that fits compression schemes with given corpora.

Admins can keep an eye on get right of entry to with SQream DB’s role-based permission machine, and the platform scales routinely with the addition of garage and compute nodes. It’s designed to paintings on any Linux and Nvidia CUDA-capable server. Rising SQream DB in both path doesn’t have an effect on knowledge availability or integrity, the corporate says.

Whilst GPU-accelerated databases like SQream are well-suited to sure workloads, it’s unclear whether or not they’ll transfer into the mainstream anytime quickly. That’s as a result of they have a tendency to accomplish poorly on database operations that may’t be parallelized or that don’t contain floating level and different numeric processing. Plus, they fight to face out — SQream competes with distributors like BlazingDB, Kinetica, and OmniSci (previously MapD), amongst others.

The corporate says it’s particularly focused on AI and information analytics programs requiring top database throughput. One buyer — AIS, certainly one of Thailand’s biggest cell phone operators — deployed SQream DB to cut back queries of thousands and thousands of uncooked knowledge information from one hour to not up to 50 seconds, SQream claims. Every other, Sheba  Most cancers Analysis Establishment, reportedly makes use of the platform to investigate as much as 1 petabyte of genome sequences.

Mangrove Capital Companions and Schusterman Circle of relatives Investments co-led this newest elevate, with participation from present traders Hanaco Project Capital, Sistema.vc, Global Business Middle Ventures, Blumberg Capital, Silvertech Ventures, and Alibaba Staff. The collection B+ brings SQream’s overall raised to over $50 million. (SQream partnered with Alibaba in 2018 to supply the previous’s GPU-accelerated database as a carrier to Alibaba Cloud shoppers.) As part of the spherical, Silvertech Ventures’ Charlie Federman and Mangrove’s Roy Saar will sign up for Tel Aviv-based SQream’s board of administrators.

Join Investment Weekly to begin your week with VB’s most sensible investment tales.


Check Also

1603674076 microsofts pivot haptics research might improve how we throw balls in vr 310x165 - Microsoft’s PIVOT haptics research might improve how we throw balls in VR

Microsoft’s PIVOT haptics research might improve how we throw balls in VR

Even though you have been separated from friends and family through time or distance, what …

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.