00:00
What exactly does a worm have in common with artificial intelligence? Well, stay tuned as we've corralled all the great new things from pure storage in one place. So let's mount up those volumes and get to it. Welcome everyone to the first episode of the pure launch round up where we bring you all the latest enhancements from across the pure storage platform.
00:28
And in this episode, we're talking about new file services enhancements and ways to make your A I workloads much faster. But first, let's talk about the Azure VM Ware solution pure storage is smashing the barriers to migrating large amounts of data to the cloud with the first ever external block storage for A VS. Now, why would you need external block storage?
00:51
Well, with A vs, all of your storage is provided from your hosts through VSAN. So if you need more storage, then you have to pay for another server to add more capacity, whether you need more compute or not. But with pure cloud block store, you can decouple your storage from your compute. Meaning that if you have large amounts of data, you can store them on pure cloud block store
01:14
which is available at a fraction of the cost of additional hosts. And best of all, if you're an evergreen one subscriber, you can leverage your existing subscription for cloud block store in Azure for all details on this and all the other enhancements mentioned here. Check out the blog in the description below. The latest flash ray purity 664 extends your file services.
01:38
So if you have more files and you know what to do with, then listen up because we're increasing the limit of files per directory from 1 million to 4 million. So if you have large data repositories such as a radiology pack system, genomic research or something as simple as log management, this opens up a ton of flexibility.
01:57
So with this four X increase, you can be more efficient in how you manage your large data repositories with up to 5 billion files on a single flash array plus with higher capacity arrays such as flash array, E, you can further eliminate the need to splitt those large directories into smaller subdirectories. And how do you increase the throughput to access all of these files?
02:20
Well, L AC P trunks obviously. However, in the past, if you needed your virtual interface for file services tagged to a specific V land for isolation, well that you were stuck making the choice between L AC P or VL A tagging, you just really couldn't have both. However, now with flash rate purity 664. We've expanded the virtual interfaces to
02:42
support L AC P with va tagging layered on top. So now you can have your traffic separation for security and link aggregation for performance too. If you're in the financial services space, then this one's for you. Flash blade, right? Ones read many capabilities ensure that your
03:00
objects cannot be deleted or modified once committed to storage. And we're happy to share that our worm capabilities are undergoing third party compliance reviews for the Securities Exchange Act section 17 A subsection four F and section 18 A subsection six E, basically, these mandate members of the National Securities Exchange must preserve their financial data for between three and six years.
03:26
And P's implementation of worm for your object storage will soon help you meet those regulatory requirements, making it a blazing fast and very easy decision for those of you who need the compliance without jumping through a bunch of hoops simply turn on object lock with worm compliance and know that your data is protected and immutable for the period that you specify. And speaking of keeping your data safe,
03:50
many customers will use replication to keep copies of their data on separate storage appliances that are often located in different locations. And if you're using flash blade, then consider upgrading the flash blade purity 440 to take advantage of improved fan in and fan out replication. For example, if you're running flash blade s or the first generation flash blade,
04:12
you can have array level connections with up to 10 other flash blades, meaning you can replicate your snapshots to as many as 10 other similar model flash blades. If you're running flash blade E, then you have up to five array level connections with other flash blade E systems. You can also have bucket level fan out replication with up to four buckets in a full
04:34
mesh deployment across different flash blades and bucket level fan in replication with up to 10 buckets for flash blade s in first generation flash blade and up to five buckets for flash blade E. And if that's not enough to get you fired up to upgrade your flash blade, then how about up to 20% more usable capacity just by upgrading purity.
04:58
Seriously, we've implemented efficiency improvements that reduce overhead. So you get more usable capacity with the same configuration and reduce your overall cost per usable terabyte. Moving on to artificial intelligence. Pure storage is a leading provider of validated proven reference architectures for A I.
05:20
In addition to res which is a certified NVIDIA DGX space pod reference architecture, we recently published two new flash stack cisco validated designs for A I. And now we're a validated storage partner for NVIDIA OVX server solutions. Now OVX servers are powered by Nvidia's L 40 S GP US for ultra high performance A I computing power. And when paired with flash blade s under this
05:49
new reference architecture, you get all the best in model training performance while maximizing your energy efficiency. Additionally, flash blade S is the perfect solution with GP U direct storage certification from NVIDIA. This provides sustained throughput that A I workloads require by eliminating CPU processing
06:10
bottlenecks. So you get unmatched performance for all your large A I workloads. And it's all part of our goal to provide you with the best full stack ready to run A I infrastructure. And with the explosion in demand for generative A I and retrieval Augmented generation, which is basically aimed at improving the accuracy of
06:32
large language model generated information without having to constantly retrain those models internally. Pier has had great success running our own rag pipeline on our all flash storage and we're not keeping this goodness to ourselves. For example, we've also developed a financial services rag to help customers summarize and query large amounts of financial documents and we're even working on new healthcare and public
06:59
sector rags as well. So there you have it all the latest and greatest hits from across the pure storage platform. Now, which new capability are you most excited about? Let us know in the comments down below and for more updates, be sure to subscribe. So you don't miss out on future episodes.
07:16
And while you're at it, give us a big thumbs up to let us know that this was useful to you. If you're looking to stay connected to pure storage and to be the first to know about all of our latest and greatest innovations. Be sure to check out the new pure launch web page by clicking the link in the description below. Thanks so much for watching and until I see you in the next episode,
07:36
stay pure.
