Migrate Cloud Block Storage – Wayne Lam, Cirrus Data
Users can now migrate block-level data from anywhere into any of Azure’s block storage offerings using the Cirrus Migrate Cloud Block storage. It delivers high performance and provides granular access to raw data, making it highly suitable for Tier-1 business critical applications. The versatility of block storage also allows it to be used as boot disks (OS disks) and data disks supporting virtual machine-based file systems, RAID arrays, databases, and container applications. On-premises, scalable block storage is mainly provided by SAN (Storage Area Network) appliances, and Cirrus Data helps migrate those to Azure – for example, to Azure Elastic SAN (in preview).
Transcript
This is Techstrong tv. Hey, everyone, welcome back here to techstrong tv. I'm really happy to introduce this next guest to you.
He's a fellow who's had a, you know what, uh, remarkable career. It continues as the co-founder CEO o of his latest company, sir Data, I want to introduce you to Wayne Lam. Hey Wayne, welcome to techstrong tv.
Hey, how are you doing? I'm doing well. Um, Wayne, I, I hope I didn't embarrass you, but you know what?
I, I already kinda let the cat out of the bag. Let tell people a little bit about your career and your journey. Sure.
No problem, Alan. Yeah, I'll be happy to do that. So, yeah, I really, I think some of my colleagues joke about it.
They say, I was born with a, uh, a piece of a storage, uh, a hot drive my tongue. But, um, yeah. So even before I graduated, uh, and I went to, uh, one of the pretty interesting, uh, very prestigious engineerings school in Manhattan called Cooper Union.
Sure. And we are known with the industry and so on. So, yep.
Long story short, in about five minutes is that I started the first company as a junior project working with the, uh, the industry. One of, uh, some investor and founded a document management system that's a scan view store print. So that was my first company founded with a team of all my classmates, uh, that went on for about three year.
And we then spinned it off and become a storage focus company. So we took scan view, store print, and focused on storage. And believe it or not, okay.
That second company called Apply Programming Technology, created the world's first object-based storage. And I'll tell you why at that time, it is warm, right? White once we many, these are basically gigantic CD PLAs on 12 inch PLAs to give you like one gigabytes of storage.
And they have a big refrigerator, robotic toolbox arm that can automatically load and feed into the storage, uh, the, the, the, the, the dry and then take it out and so on. That gives you about 50 gigabytes of storage. Wow.
Right. You, you know, you can hold it in your thumb, thumb dry nowadays, but back in 19, now you Can No, I remember those. Yep.
I remember the hard cards. Me 40 meg hard cards. Absolutely.
So, so those are the good old days. So, so, so we created, because Worm is not really lend to sell to writeable operating system or file system. So we created object storage.
You gimme an object, whatever it is, I stored it and I give you back a handle. You what they now call object id. Right.
Okay. Right. When you gimme back the id, I'll give you a file, that's it.
And we put in a d i and so on, and that become pretty successful in the document management, uh, uh, um, um, market at the time. And we license into some company, and one company took, took interest of us called Cheyenne Software. So if you recall, back in the old day, right, Cheyenne, I remember them well.
So they call me and say, Hey, Wayne, I wanna buy your company. I said, really? Why?
Why you think I'm for sale? Right? But, but then I end up taking the deal.
We, I joined Cheyenne as employee one 60. We grew the company to 2000 people. Right.
And worldwide with, you know, hundreds of million dollars of revenue. And then Charles won. So that's my second, um, actually one, two, that's my third.
Right. And then Charles Wong, computer associate board. Yeah.
3 billion. Okay. Wow.
That's actually, I Remember that too, now that you mention It. Right, exactly. So that's my, uh, uh, uh, third or four.
I lost KGO already. Yeah. So I worked for ca Golden Hancock, what, whatnot.
Right? So Charles knows that I'm not a, uh, uh, um, uh, somebody worked for 12, 1200%, uh, uh, 12,000 people, uh, company. So he know that I'm gonna do the next thing.
And sure enough, after three years, I spin off and I become, uh, a f co-founder Falcon. So, so we do the world's first virtual storage thing, storage over ip, right? Snapshot.
Mm-hmm. Uh, and at the peak of, uh, of that company, we actually actually licensed the virtual tape library to EMC at the time. Right.
And that was very successful. So all the virtual tape drive out there by EMC is, is actually our OEM product. So did that for almost 10 year.
We founded out in, uh, the year 2000, right. During the ba that bomb days, it went public, you know that, right? So then, you know, at the, in, uh, 2000, uh, um, 11, I found that that's enough.
And due to some other events, I found myself, you know, really better off, you know, on my own again. So I found the service data, and I've been running it since then, right after almost, uh, uh, 12 years now. Wow.
Great. Yeah. And, and so it, that's kind of my, my career.
And the good thing about this group that I have is that they follow me through all these five, you know, four or five different ventures. Uh, very, very competent set, very elite, and very, very focused on storage. So we have so much storage equity from fiber channel protocol, target mode, initiative mode.
I like. Right? Anything block.
So when I left, uh, uh, um, a Falcon store, the thing that I want to do is to say, you know, all this storage virtualization stuff is great, but how do we make it better? And the biggest pain point in adopting a events device between the host and the storage and do something useful in the middle, is how do you insert it into the data path without any outage? And we figured it out.
So we patented this called Transparent Data path interceptor, and we have like 11 patents surrounding that to defend it, to make sure every end goes clever, okay? Mm-hmm. In short, allow you to do this, I can, so the first use case we said to test this thing on is migration.
And we got Right. You know, and we're in there and we never got out of it. We have so much things we want to do with this technology.
So in short, what it can do is that, let's say you're running an Oracle workload, Oracle wax asm, right? 10, 20, no, you know, gigabytes of performance and so on, really busy, zero downtime, and you bought some new storage, new fiber channel storage, and you want to go from A to B, but you have no downtime. My technology at the net net of it was the, was was the really compelling reason why people want us, okay?
Is that I can migrate all that Oracle database from A to B cut over running on that transition, the whole thing. And the host owner doesn't even know that we did his Oracle keeps running and all sudden, oh, wait a minute. Why is the Oracle not running five times faster?
Well, you just migrated to all flash, okay? And serious data did it, and it's so serious, serious data, sorry, did it, and it's so transparent and zero impact that nobody even noticed we did it. That is, if you can imagine if I can deliver this for the storage vendor, they love it, right?
Why? Because it's such short the entire, uh, sale deploy. And, and how do you then transfer the old storage from your competitor to you, right?
In a smooth way so that the customer can sign off, right? If you don't do that, the storage is sit on the shelf waiting for downtime to migrate the data over, which is often the case even today, right? Without, without our solution, right?
So we, we solve a big problem for all the storage vendors, so all the tier one vendor, right? We talk about Dell, right? No more emc, Dell emc, right?
Dell, you know, uh, pure storage, right? H p e, you know, I bm all, you know, NetApp, all the, uh, h hadachi, sorry, gotta mentioned them all the storage vendor, the tier one vendor, are using our product one way or the other. In the back end.
We had the, we had the, you know, we had the unsung hero for up to now last count one exabyte, one exercise lot data have been migrated by us, and nobody know that we existed until now. I have this opportunity to be on your TV and talk about It. What, what a what a great, what a great story, man.
Fantastic. You know, listening to you talk, it's kind. Uh, I kind like, you know, the times of my life, right?
You know, that song, the Times of my Life and, and brings back so many memories and stuff. And you know, as we were talking off camera for a long time, Wayne, or not in a long time, but recently, we've, we've focused so much on the application mm-hmm. And, and the interface and, you know, the cloud and setting up our instances, and the only time you would hear about data is when people would frankly b***h and moan about how much it costs to store data.
And do I really need to store all of this data? My data costs are exploding, right? Cause I have so much more data now.
Mm-hmm. And, and I have to, you know, and I'm using ML and I'm using all these technologies to, you know, analyze it and, and all of that. But I, I think I, I saw it kind of with C O V I D as we kind of started emerging out of covid, people began to realize that that data, those are the crown jewels that is worthy of our money, right?
And the other thing is, I think we began to look at data through new lenses, if you will, where not all data is the same. Not all data is equal. Some data probably, or not, probably some data is more valuable than other data.
Mm-hmm. And, you know, with, with new technologies, we, we can be better and smarter about segmenting our data. Do I want this on flash?
Because it's much, much faster, and it's the kind of stuff that I'm accessing pretty regularly. So I want, I, I need that la you know, I need that speed where this kind of data, I'm more archiving. I might hit it when I need it once a year or something.
A little bit more latency there is not fatal to the equation, right? Yep. And so, being smart about segmenting my data is, has become a, a real, it's a business.
Mm-hmm. Yes. Right?
It, it's a business. And it's, and I imagine, you know, part of what you are doing at, at Sirius is, is helping people be smarter Yes. About their data.
And then, you know, it's part of that migration, right? We don't just migrate the whole mass, if you will, into Yep. Colin mm-hmm.
The next place. Yep. Now, to me, this is something that calls out for ai.
Mm-hmm. Right? Well, how can AI help us with that?
Yep. And I'm sure you're on top of it because you've been on top of it for 30 years. Absolutely.
Yes. So, so AI can certainly play a very interesting and, and, uh, useful role in the aspect of, you know, so focusing on, on, on, on my, uh, market here, which is for storage migration, right? I mean, just to make the decision of, okay, so this is my set of disc, and as you said it, not all disc are equal, right?
So even in a modern, uh, uh, array on Preem, they have different class of a storage inside or flash, some older ones still have spinning disc, believe it or not, okay. SAS and so on. So, and, and it's different cause for the performance, uh, uh, that you get, right?
So what, which makes more sense. So yes. So now everybody is talking about migration to cloud, right?
So after all this one exercise of migration that we've done on-prem for the storage vendor on-prem, we of course, uh, about a, a year ago relaunch into the cloud, uh, our ourself, okay? Cloud is really the next thing to go. And in cloud, if you see the on-prem, the different tiers of storage required for different types of data, use, use cases when the cloud is even more so, right?
So a w s Azure, you know, Google, they offer different tiers of storage, and they are drastically different in cost per terabyte, per month. The cheap one, they don't perform so well, right? And, but they're very low cost, you know, Penn is, uh, per gigabyte per month, but the higher one cost is dollars, you know, or even more per gigabyte per month.
So a terabyte comes up to thousands of dollars per month, but they give you the amazing performance. Like, not, not, not even the on-prem can match. So the key is how do I balance when I, when I'm migrating a, a a hundred terabyte from on-prem now to the cloud, if you just say, okay, just allocate a a hundred terabyte in the cloud without really going through the analysis, you'll be raising a lot of money, buying excess, uh, performance that you don't need.
Or worse the other way around, you did not buy enough the performance. Now your Oracle runs like, like a doc, okay? And neither, neither case is good.
So it's all a matter of right sizing. So most recently we launched with a, sorry, take back that word, not launch, but we collaborated with, uh, Microsoft Azure, right? And we actually build a storage analyzer that optimizes, that can analyze, look at your source performance requirement, actually monitor it for a period of time, right?
Because without realistic data, it's garbage in, garbage out. Okay? You make too much assumption, okay?
So we allow actual measurements of your host, uh, uh, workload in terms of the performance, and will be very accurately mapping each of the tiers of source storage into the cloud. And some of them might be by mistake. Wow, you put this in N V M E and yet it only generated five I up per second, you don't need that.
Go to the cloud. Now you can save a whole bunch. Sure.
Right? Instead of matching it. So with this, we'll be able to help the customer rightsize the solution as they migrate to Azure.
And, uh, I think Microsoft is launching the, the elastic send, which is another tier of, of more cost effective, uh, uh, um, kind of like a virtual array in the cloud kind of a concept, right? And, and will be with them, and we'll be, uh, launching our capability with them to help their customer rightsize the solution. Excellent.
Yeah. So that's that. I'm, I forgot to close this thing on the ai.
So in this thing, our initial thing is just based on actual performance, but AI role will allow me to trend. For example, performance is a constant, this one is a hundred thousand iop, but it's probably only active during four o'clock in the afternoon for some reason. Okay?
And so with AI involved, it will make a more dynamic assessment rather than a static assessment, right? Looking over a one week period. And who knows, maybe for some reason when this workload peaks here, two hours later, this workload would peak for whatever reason.
But AI would know, okay. Right? And it would be able to forecast some of those things so that in the picture, I think as we evolve our capability, it can adopt some of these capability, uh, these, uh, uh, advancement in AI to help more accurately, uh, make the, uh, assessment for the cousin.
Love it. Um, so that, that's big, big news with Azure. And of course that's using whether you want to call it AI or ml, right?
It, it's using intelligence there to, to help you do this sort of in an automated fashion. Also, some news, you mentioned Dell, Dell emc, um, recent news with Dell there also using sir. Now, is that more for the, for on-prem storage type of stuff, or, Yes.
Mm-hmm. Yes. So I think the, generally the industry analysts in the storage space who focus on the space generally agree that their waves of migration to cloud, right?
So let, let's, let's all, you know, recognize the cloud is one of the most important topic, right? So because it does provide, when used properly, when migrated properly, it can help you, you know, check, uh, um, uh, manage your source and transform some of your op, uh, uh, CapEx expense into the opex, which every ccf, right? Love to engage in a conversation just by hearing that alone, okay?
So, so cloud makes sense, okay? But you have to do it, right? Right.
So even in the cloud, the old days of just having the native cloud block storage okay, is over, right? Even so, Azure has taken the lead, uh, uh, the, the position to say we are gonna create our own virtual array. So, you know, instead of managing, allocating, sorry, allocating and managing in that order, a instead of disc block storage for your Oracle, for your SQL and so on, right?
Why not manage them, manage them as a virtual storage array, because that's how the on-prem world works, right? On-prem. You don't go to come USA if they still exist.
No, they don't exist. Uh, No, they don't exist anymore. But I used to go to the one on Long Island.
I I remember Windows 95 was released. You midnight? I'm online there, I go to Michael Center now.
Yes. So I go, I go sensor in, uh, uh, uh, Meadowbrook, uh, uh, pub, okay? And they're still there.
Yep. You don't go and buy one terabytes, another one terabytes, and, and, and then you manage it like that, right? You, you put into a storage array, right?
Where you can have protection, Snapchat and virtual storage and thin storage. So, so Azure is, I think, the first known event cloud vendor to say we are gonna provide that instead of distinctive lung storage. So there are a lot of management cost saving, something called thin provisioning, right?
The customer love to hear thin provisioning at a glance in, in, in one sentence is just in time paying, okay? For the storage I want, right? So, so in the old days before, before the virtual array existed, when you license one terabyte, even before you copy a single byte, so that this is completely blank, how much you pay for per, per, per month per gig, the entire one terabyte.
Okay? They don't care because they don't know, right? But there's a virtual array, I'll be able to create virtual this that looks like a terabyte, but before I write to it, it doesn't allocate any storage.
Now, that's really cool. The customer can enjoy a lot of savings. Okay?
So, so that concept is good, along with faster snapshot, better snapshot, better data protection. So of course, the storage vendor that are OnPrem, right? For example, pure now have their cloud block source c b s, that is a virtual array, right?
NetApp was really one of the first, uh, to have done that Yeah. Many years ago with the C V O, right? Uh, and then H P E and everybody else, uh, followed suit with that, right?
And the latest relationship that we have, which is an amazingly great relationship, is with Dell, right? With the Apex, okay? So Dell has this whole new initiative launched just by, uh, uh, a couple months ago in their applic kickoff in Las Vegas, okay?
The push for cloud, push for Power Flex, which is the new product offering in the cloud that's completely virtual and so on, that offers a number of benefit. And we are a partner of theirs. And I think, to the best of my knowledge, we are the only solution that can migrate from traditional cloud block storage to the ap, to the apex, to the, to the PowerMax, right?
Uh, so, so all these things ref, uh, re um, reflects very innovative, uh, uh, uh, uh, solution that we have Okay. To transition to help the customer transition from on-prem to the, to the, uh, to the cloud. Yeah.
And I'm very happy to be part of those. Yes, Absolutely. Wow.
Wayne, we're, we're over time already, but I didn't wanna stop you, right? It was, it was interesting stuff for people who, you know, who are saying, Hey, I'd like to find out more about Sirius. I may not be Dell, I may not be, you know, Microsoft, but, you know, storage is important to us.
We have a lot of data needs. I'd like to find out more about what, how I can get help with migration. What's their best path?
Absolutely. com. Okay?
Yep. It's service is spelled C I R R U S as in the service cloud, cloud. I'm a pilot, that's the name of my, uh, airplane.
Okay. And really, oh, the company. I've been in the name of the airplane, like Cheyenne Falcon, you know, is the Falcon Jet.
Yeah. Cheyenne is the, so now Serious. Yeah.
And now serious, right? So, yeah. So go to our website, okay.
And, uh, uh, um, and, and we have a row of, uh, uh, uh, white papers and, and, and, and video to demonstrate how to, how, how things are done. And on our mi our cloud offering is a migration as a service. So it's really a platform that allow anyone for free to go there and sign up for an account, and we will give you some free license, and you can migrate your local drive to the cloud, or from cloud to cloud, or all those combinations.
So I encourage people to do that. Okay. And yeah, and there's a lot of great articles published and by, you know, after interviewing with, uh, leading journalists like yourself, okay, that explains why we did it, how we, you know, deliver real benefit for the customer.
So we welcome everybody to go to our site intake club. Fantastic. Wayne, thank you for coming on and, and educating our audience a little bit today, making 'em a little smarter about storage.
You know, as I mentioned, we, we'd love to have you back on. Whether it be on these shorter interviews or longer ones, or some of the events we do, you're, you're a treasure. So thank you very much, and thank you for all you do.
You're you're welcome, Alan, and thank you for having me. All right, Wayne Lam, co-founder, CEO O Sirius Data here on Textron tv. We're gonna take a break.
We'll be back in a moment.