SQream DB

The increasing volume, velocity and variety of data is driving the need for more efficient data management. This is accelerating the growth of GPGPU (General-purpose computing on graphics processing unit) computing as an alternative to CPUs via the manycore route. GPUs today offer thousands of cores versus the tens of cores of CPUs. This factor, along with a simpler software model (CUDA/OpenCL) when compared to CPU based multi-threading, makes GPGPUs a superior alternative to traditional CPU compute models. GPGPUs have already demonstrated their value in high performance computing for over a decade. Now they are finally moving into enterprise computing. For the first time, there is a ground-up implementation of a mainstream enterprise RDBMS on a GPU - SQream DB. Marlabs is the strategic partner for SQream Technologies in the US, helping clients implement high ROI business solutions using SQream DB.

This flagship product from SQream Technologies provides organizations with the most rapid, cost-effective, petabyte-scale SQL database available in the market today. Many benefits of GPGPUs, thousands of cores and low-level software are kept reliably wrapped in a relational engine, allowing developers to reap the full benefit of underlying GPUs. 

SQream's huge accomplishment besides performance is the tremendous business value it delivers in making ‘no change required’ by developers, power-users and end-users to use GPGPU. SQream users can be completely oblivious of core counts, CUDA, blocks, warps, etc., but still benefit from all these technologies in a seamless manner with good old SQL. Even a mainframe developer would feel at home with SQream. Simply put, SQream is the ‘iPhone’of relational databases—a true paradigm shift.

Key Benefits

  • Real-time rapid insights
  • Single-point-of-knowledge
  • High ROI
  • Low TCO
  • Optimized for petabyte
  • Rapid implementation
  • Easy and intuitive usage
  • Maximization of existing core IT investments

SQream in Action

Cellcom, Israel's largest telecommunications company, used SQream DB for software network optimization—analyzing CDR rollup and general performance event handling (GPEH). Run time for the preparation cycle involving CDRs download, mediation and insert reduced from 30-40 minutes to 4 minutes. Storage size reduced from 500GB to 100GB without indexes using SQream DB. Formerly unused diagnostic logs transformed into valuable insights on network activity, which were physically unattainable with the previous infrastructure.

A medical center used SQream DB for genome analysis of DNA data. The work involved loading and analyzing massive quantities of BAM files—simultaneous querying of 50 BAM files, 30 billion rows of sequenced data with drill-down to specific position in chromosome. This resulted in a first-time ever presented ability, an effective data modeling for SQL querying of BAMs. The time reduced from 1 month using manual to 60 seconds using SQream DB.

- Less

SQream DB

The increasing volume, velocity and variety of data is driving the need for more efficient data management. This is accelerating the growth of GPGPU (General-purpose computing on graphics processing unit) computing as an alternative to CPUs via the manycore route. GPUs today offer thousands of cores versus the tens of cores of CPUs. This factor, along with a simpler software model (CUDA/OpenCL) when compared to CPU based multi-threading, makes GPGPUs a superior alternative to traditional CPU compute models. GPGPUs have already demonstrated their value in high performance computing for over a decade. Now they are finally moving into enterprise computing. For the first time, there is a ground-up implementation of a mainstream enterprise RDBMS on a GPU - SQream DB. Marlabs is the strategic partner for SQream Technologies in the US, helping clients implement high ROI business solutions using SQream DB.

This flagship product from SQream Technologies provides organizations with the most rapid, cost-effective, petabyte-scale SQL database available in the market today. Many benefits of GPGPUs, thousands of cores and low-level software are kept reliably wrapped in a relational engine, allowing developers to reap the full benefit of underlying GPUs. 

SQream's huge accomplishment besides performance is the tremendous business value it delivers in making ‘no change required’ by developers, power-users and end-users to use GPGPU. SQream users can be completely oblivious of core counts, CUDA, blocks, warps, etc., but still benefit from all these technologies in a seamless manner with good old SQL. Even a mainframe developer would feel at home with SQream. Simply put, SQream is the ‘iPhone’of relational databases—a true paradigm shift.

Key Benefits

  • Real-time rapid insights
  • Single-point-of-knowledge
  • High ROI
  • Low TCO
  • Optimized for petabyte
  • Rapid implementation
  • Easy and intuitive usage
  • Maximization of existing core IT investments
+ More

GenomeStack is an innovative platform developed specifically for bioinformaticians. A new concept in the genomic research industry, GenomeStack provides researchers with the ability to store and effectively query massive amounts of genomic data such as BAM, VCF, fasta, and fastq files on a columnar, big data SQL database. It delivers insights up to 100 times faster than any other available solution. GenomeStack is a hassle-free, easy-to-use and implement solution that will enable bioinformaticians to focus on the genome research instead of the computational environment. 

With GenomeStack, post-aligned sequencing data from multiple samples is stored in an ultra-fast and highly scalable database, enabling the querying of a large number of samples simultaneously. The power of GPU when encapsulated in commodity servers delivers significant savings in hardware, power consumption, cooling systems and floor space. When combined with the SQream database, there are also significant savings in storage, compared with other storage solutions that require full floor-to-ceiling racks. With GenomStack, bioinformaticians benefit from the power of a high-speed GPU-based columnar database, leveraged by a high-capacity storage management system, with transparent genome-specific compression capabilities.

Read more

xDR Repository is a big data SQL database platform developed specifically for the telecommunications industry. As the digital revolution gains pace, and telecom service providers get flooded with big data constantly generated from multiple sources, telecoms are faced with a growing challenge. xDR Repository empowers engineers, data scientists, analysts and marketers in the telecom industry to extract value from their ever-growing data volume. 

xDR is a rapid and effective platform for making accurate predictions and reaching intelligent decisions—working directly from a single point of knowledge, gathered by an ultra-rapid, large-capacity, petabyte scalable and cost-effective solution. With xDR Repository telecoms are now able to enjoy the strengths of a high-speed GPU-based columnar database, exhibiting enormous scalability capabilities—leveraged by a high-capacity NAS dataset management storage system, with transparent CDR-specific compression capabilities. 

Read more