Reference architectures around the synergies between Intel Optane persistent memory and Alluxio’s open source data orchestration software for the cloud using in-memory cache will initially sell through the AWS and GCP marketplaces. A rollout to a specialized integrator channel, and to Azure will happen during 2020.
Today, cloud data orchestration provider Alluxio is announcing a joint Go-to-Market solution with Intel that brings together Alluxio’s in-memory data acceleration layer with Intel’s 2nd generation Xeon Scalable processors and Intel Optane persistent memory. The channel play, being boosted by Intel introducing the solution to its own key partners, will start with cloud marketplaces and broaden out from there.
Alluxio is a commercialization of the Tachyon open-source virtual distributed file system, which was originally the cache and file system for Apache Spark, and was part of that distribution incubated at UC Berkeley’s AMPLab.
“We were told eight years ago we needed to bring compute and storage together to deal with constraints, but when we did that, there were different constraints, including competition for resources,” said Rowan Scranage, Alluxio’s Chief Business Officer. “Our goal at Alluxio is to accelerate the use of compute frameworks leveraging machine learning and analytical applications, and to do that, we do two things. We cache data across any systems, making it available in memory so you don’t have to go back to the storage. And with our common areas, we pull data from all areas and present it back as if it is native, regardless of where it is stored.”
Scranage said that Alluxio customers use their acceleration to burst to the cloud, even when some of their compute is on premises. Others use it to make migration to the cloud quicker.”
Many of their customers are large, including WalMart, Rakuten, Barclays and eight of the top ten Internet providers in China. Not all are, however.
“We also do a lot of work with startups, including a number of Israeli companies,” Scranage said. “They can work with us on the open source element and migrate to our enterprise version over time. Some of these have revenues of sub $5 million, and some are even in the pre-revenue stage on open source.”
Their initial focus in terms of partners were the big cloud hyperscalers.
“We have strong partnerships with AWS and GCP, and are growing with Azure,” Scranage said. “We put those strategic cloud channels in place before going out to distribution. A bad channel program is worse than no program now. We are now at the stage where we are looking for more boutique systems integrators and analytics specialists. Many of these are local cloud partners for GCP and AWS.”
The new Go-to-Market announcement around Optane follows an initial technology integration between the two companies.
“The initial integration was baby steps, getting a good working relationship in place with Intel,” Scranage said. “We have joint webinars on the solution already, and other pieces that will come forward in the coming months.”
This announcement is all about bring the collaboration to market effectively, to help companies with their challenges around bounded storage and compute resources.
“This deals with the issue of customers’ need to spend less on compute and storage,” Scranage said. “If they were already using us, they will get better performance. Through this partnership with Intel, we provide customers with guidance on reference architectures that give high performance. We do joint testing and create these reference materials. Intel is also introducing us to their own channels, making their channel more aware of us.”
In addition, Alluxio has been talking with Intel’s own internal IT, which has been taking Alluxio’s sales teams into Intel’s customers. Alluxio has also joined the Intel Software Innovator program.
“It’s a credit to Intel for the way that they partner with smaller organizations like us,” Scranage stated.
Benchmarking results from the collaboration show 2.1x faster completion for decision support queries when adding Alluxio and Intel Optane persistent memory [PMem], compared to only using disaggregated S3 object storage. Using Storage over App Direct, a feature of PMem App Direct mode, lets Alluxio access high-performance block storage without the latency of moving data across the I/O bus to provide the data acceleration and reduction in query runtime. Accordingly, with Alluxio and Intel Xeon Scalable processors, an I/O intensive benchmark delivers a 3.4x speedup over disaggregated S3 object storage and a 1.3x speedup over a co-located compute and storage architecture.
The Go-to-Market channel strategy with Intel will initially focus on the AWS and GCP cloud marketplaces, but that will broaden out during 2020.
“We are in those cloud marketplaces now,” Scranage said. “Our own direct customers who are looking for recommendations in reference architectures will be provided with those. We are not yet in the Azure marketplace, but that’s a goal for 2020. Another goal this year is to add data analytics boutique Sis. DBS in Singapore, which have been acquired by Accenture, is a good example.
Scranage noted that they have seen good interest from a number of Canadian companies.
“That has been organic, because we have been quiet there from a marketing perspective,” he said. “Some of the communications companies in Canada have reached out to us.”