…and 5 Facts to Back That Up
I started a new job at WEKA late last year and received a lot of compliments here and IRL. But every so often I would get a message or be pulled into a conversation that mentioned me working at yet another storage company which led to a conversation about what WEKA actually helps customers do. I’ve responded in private quite a few times, so I decided to write a post about it.
Let’s get this straight: While I am proud of the kick ass work I have done at storage companies over the last 20 years, I am not working at another storage company! WEKA is a Data Platform company and one focused on AI and other modern, high performance workloads. And no, “Data Platform” is not marketing buzzwords meant to make old technology seem more appealing. We really are very different from the storage companies I have worked at. And here are 5 reasons why!
#1 First and foremost, we don’t sell disks.
WEKA sells software as a subscription that is highly scalable and easy to deploy, configure, manage, and expand – be it on-premises or in the cloud. On-premises, we are hardware agnostic and able to run on almost every hardware out there – from multiple server vendors, with multiple types of NVMEs drive, using multiple CPU choices, feeding data to multiple types of GPUs, and on multiple clouds. The WEKA Data Platform runs on servers from our customers’ preferred vendors and can be bought in bundles from them too. This makes ordering and managing it much simpler than a separate storage appliance. Our software can be deployed on standalone server nodes or for zero extra footprint in a “converged” configuration on the servers that run the workloads.
#2 We are a cloud software company – who also run on-premises.
I mentioned cloud above, but it deserves a full call out. The WEKA Data Platform was designed to be cloud first and then ported to on-premises. And as a result, it allows for total data portability and management across multiple environments – cloud or on-premises or both. Try getting the same <insert legacy storage array name here> experience in AWS and on premises. And for bonus points, try enabling hybrid workflows between them! WEKA gives customers a consistent data platform across any environment, be it on premises, in the cloud or both. And too often “hybrid storage” means limited cloud capabilities like snapshots or backup. WEKA customers are using the data platform for a plethora of cloud uses – bursting from on-premises to the cloud as needed, supporting sophisticated hybrid AI workflows, and migrating workloads to AWS Outposts or the AWS cloud.
#3 We treat all types of workloads equally – blazingly fast.
Storage systems are usually architected to excel in a limited number of performance dimensions while others languish, requiring different storage arrays for different workloads. But WEKA performs across all dimensions without the need for tuning or re-configuration, customers can run any part of their data pipeline on a single system, whether it requires massive IOPS with small reads and writes, or massive throughput and 10’s to 100’s of GB/sec. But it’s not the speed that is important – it’s what it means for our customers. It allows them to have just one storage system for the entire data pipeline and save money from having to buy and manage multiple storage arrays and equally as important saving time from constantly copying data between them.
#4) We evolve as fast as our customers’ needs do.
Traditional infrastructure can take months to years to change. New hardware support or networking protocols are often only implemented in storage arrays years after they are released to the market. And to get new features you often have to undergo painful data migrations and hardware swaps. We don’t lock customers into specific hardware platforms or previous generations of technology. Platforms are designed to be future proof to be extensible and evolve with customer needs to take advantage of new innovations so their business and research can benefit. The WEKA Data Platform is designed to evolve at the speed of science to let customers quickly, easily and non-disruptively adopt new technology.
#5) We are faster than a storage array and we are even faster than local NVMe storage – by up to 3X!
Local storage means local resources only. We parallelize and gang together more NVMe resources than can fit in a single chassis, add a high performance networking stack as well as kernel bypass to service high performance workloads faster than local storage. A WEKA cluster can deliver up to 162GB/Sec & 1M IOPs to a single GPU client. Earlier this year, Samsung Datacenter Technology and Cloud Solutions used WEKA to bring their NVMe drives together in a storage cluster and achieved new number 1 records for 4 of the 5 SPEC Storage2020 audited benchmarks, all using an identical configuration across varied datasets and IO types. Yes, it’s a storage benchmark, but until a Data Platform benchmark is created it’s what we have to work with.
I am very excited for my next, non-storage adventure at WEKA. If you want to learn more about the data platform, why not try it in AWS right now by going to https://start.weka.io. In addition to all the great benefits we bring to customers, I am working with a team of the best-of-the best from across our industry. And I am going to be hiring product marketers and technical marketers VERY soon so stay tuned for job postings.