In today’s data-driven world, enterprises are constantly faced with the challenge of managing and processing vast amounts of data in a timely and efficient manner. With IBM formally announcing the end-of-life for IBM’s Mako-generation Netezza systems, many companies are now forced to consider their alternatives as we count down the final months left for vendor support.
Yellowbrick’s Heather Brodbeck, VP Revenue Operations and Umair Waheed, Head of Product Marketing are joined by Asim Aziz, CTO of JMA Resources and Huw Ringer, CEO of Smart Associates to discuss the alternatives for organizations and how to navigate these changes.
The panel discusses the importance of modernizing your enterprise data warehouse to keep up with the changing data management landscape. They also highlight the benefits of migrating from IBM Netezza to other options that offer better performance, scalability, and flexibility. With insights from real-world use cases and customer success stories, this video provides valuable information for enterprises looking to optimize their data warehousing capabilities.
Transcript:
Heather Brodbeck:
All right, here we are. Thank you everyone for joining our LinkedIn Live event today. My name’s Heather Brodbeck. I’m the VP of Rev Ops at Yellowbrick, so I’ll be hosting our panel today. We’ve got a great topic, “Sunrise on Your Enterprise Data Warehouse, Sunset for IBM Netezza.”
We’re going to be hearing from some experts that have a lot of experience, both working with Netezza, working with companies to migrate off Netezza. We’re going to go over some information on how to go about that, what the options are, what the cost is, a lot of different topics. Thanks for joining. And so, I’ll just go through our panelists here as a start.
First, we have Asim Aziz. He’s a CTO of JMA Resources. He comes with expertise in architecting solutions, implementing technologies for data warehousing, data science, multitude of different data-centric capabilities on a global basis. He is a senior member of the Institution of Electrical and Electronics Engineers and is an alum of MIT Sloan School of Management. Welcome, Asim. Would you like to say anything?
Asim Aziz:
Thank you very much for your kind introduction, Heather. I’m looking forward to sharing what my team and I have learned during this process with your panelists and with your audience. Really appreciate it. Thanks.
Heather Brodbeck:
Okay. All right, great. Thanks. Next, our second person joining, we have Huw Ringer. He’s the CEO of Smart Associates, comes with 30 years of experience in data warehousing, a lot of specialty in Netezza, which is why he’s joining us today. Would you like to say anything, Huw?
Huw Ringer:
Well, thank you very much for having me. And much like Asim, I look forward to sharing our experiences and learnings with your audience over that 30-year period.
Heather Brodbeck:
Great, thanks. And then finally, we have Umair Waheed. He is a colleague of mine at Yellowbrick, VP of Product Marketing, over 20 years’ experience. Also in data warehousing and analytics solutions. Umair, would you like to introduce yourself?
Umair Waheed:
Thanks, Heather. Probably coming up on 30 soon. I’d like to keep it in the 20 somethings for now, but yeah. No, I’ve been at Yellowbrick for about a year now. I’ve spent the last 25 years or so in data warehousing analytics, both on the consulting side and working for vendors. Super excited on this topic. It’s something we’ve known for a while that Yellowbrick really is a great solution for customers of Netezza. We’ve seen a lot of Netezza customers migrate to Yellowbrick throughout the last few years, really since our inception. Yellowbrick has really takes some of that great greatness that Netezza brought to the market, that innovation that they brought to the market and really takes it to that next level, but in a familiar form factor, which we’ll hear today from some of our speakers. Really excited to join.
Heather Brodbeck:
Okay. All right. Thanks Umair. All right. We’ll get into our first topic. For many people in the data industry, Netezza is a known solution through IBM, meeting a lot of advanced needs over many, many years. And so with the announcement that they’re phasing out their support, many enterprises have to make hard decisions on how they’re going to move forward. Maybe Huw, we can start with you. Can you maybe explain some of the factors that were in play for ceasing support for Netezza with IBM? Is there a concrete end of life day and where do we stand for phasing out support?
Huw Ringer:
Well, I think there were a number of factors that led up to IBM announcing the end of life for the original Mako Generation platform. Partly, I think it was due to them selling off their X 86 business to Lenovo, wanting to move all of their turn onto the power architecture. I think partly also they wanted to try to consolidate all their database customers onto the DB 2 platform.
And I think also partly, they didn’t want to have to spend money on developing two different database products simultaneously. I think those were some of the reasons and for them originally canceling the teaser. But in terms of where we are now, the end of life is rapidly approaching. Actually, I think it’s, I don’t know, something like under 70 days away, end of March-ish, early April. All those customers on the Mako Generation platform have to decide really between now and then what it is they want to do.
And IBM has announced a new version of Netezza, which they’re quite keen to encourage customers to migrate to, but the problem is that a lot of customers have been quite badly burnt as a result of this experience. They tried moving to DB 2, they found that either it didn’t work for them, or it didn’t have the same advantages that Netezza had, and they don’t want to end up in the same situation that they’ve been in before in a couple of years’ time where they’re suddenly let down again.
Hence, the reason why I think this topic is timely because people are, I think, looking for alternatives. They’re looking for solutions. They know they don’t have much time. They’ve lost trust, I think, in IBM, and they need some hope, I think, for the future.
Heather Brodbeck:
And maybe I heard you reference DB 2 in there. Asim, do you want to share anything around your experience of switching to DB 2 and maybe also give a little background of what customer or company that you were doing that with?
Asim Aziz:
Sure. We have large enterprise customers like US Navy. In this case, Huw mentioned trust. I think trust is of course basic, but it’s more than trust for us. From my customer perspective, we have followed everything that IBM asked us. They asked us to go to this company, I mean this product, and then they said, “We are going to take you to DB 2.” We took that, we bought DB 2, and then we decided, oh, we’re going to go back to Netezza. We took that.
IBM offered DOD IBM cloud platform data services, and we had really bad experience from support perspective to the point where we are not getting support that we need for general things. It is taking days, weeks to do patches for us. Something is missing in that strategy from IBM and that’s what forced us to look for something else. And then of course, there was this end of life for DOD customers. I think our end of life is somewhere around June. We had really very little time to figure out what we are going to do. And so I’m sure other customers are in the same boat like we are.
Heather Brodbeck:
Yeah. Maybe Umair, you could expand on that a little bit from your experience. I know at Yellowbrick, we’ve definitely helped some customers move off of Netezza and obviously, working with different prospects that are in that position now to try to give them support. Could you expand at all on maybe some of the challenges that you’re seeing that are either consistent or in addition to what Asim just went over?
Umair Waheed:
Yeah, absolutely. And I think the assumption that maybe IBM and others made is that all the things that they missed was the value that solutions like Netezza bring. I don’t see Netezza as just being purely a legacy technology that people have, that the laggards have not moved on from, they’re exceptions to the rule. Netezza brought a lot of innovations to the market and was highly valuable, delivered on highly valuable use cases and data exploitation for the companies that deployed it.
And so where Yellowbrick has really been successful is delivering that same innovation in not too disruptive form factor. There’s a big pressure on migration to the cloud, and Yellowbrick is certainly in that space. We have a cloud dataware offering, but also the ability to deploy a similar solution to Netezza on-premises in your data center and then have options to move to the cloud. On-prem to on-prem to cloud type methodology has been quite interesting for a lot of our customers who have a long-term vision for the cloud that maybe today aren’t quite ready or have other systems.
The data warehouse doesn’t live in isolation. You’ve got all the connective tissue, all the things that connect to that data warehouse that also need to move along with it. It can be quite a complex program. And so Yellowbrick really does provide quite a neat slot in replacement for Netezza where it currently resides in a data center or on a hosted platform and then gives people options to move to the cloud.
Now, our customers have seen the same value, improved another step change in query performance similar to what happened when Netezza came in. They’re able to deliver much more insights much more quickly with predictable numbers, but also shrinking the size of the platform. Some of these legacy technologies – they were really innovative in their time, but things have moved on and now we’re able to deliver that same experience, that same level of performance, interactive query response in a much smaller form factor, sometimes as much as 10 or 20 x smaller, running in five or 10% of the footprint of space.
And that also has advantage from a cost and energy savings perspective for those that are looking to either free up space in their data center or reduce the cost of running in their data center. And that applies in the cloud too. When we run in the cloud, we run on much smaller form practice.
Heather Brodbeck:
Okay. Thanks, Umair. And I think that leads into part of our next topic, and I know we centered this next part of the discussion around the IBM Cloud Pak option, but Huw, maybe you could share some commentary around, like Umair was saying, costs associated to that, the planning that’s required, and any risks around migration, and maybe not just with the IBM Cloud Pak option, but to other platforms too.
Huw Ringer:
Sure. Obviously, there’s always time and cost and risk associated with any migration. And so in our experience, what we’ve found is that trying to automate as much of the process as possible helps mitigate those factors really. We have created some solutions to automate a lot of that, the effort involved in doing the physical database migration and then validating that all the data matches in the source and target. We also provide an afterlife support service for customers that perhaps aren’t ready yet to migrate. They need to buy themselves some time to explore their options and think about a plan. Plan for the transition from the current environment to the new. And so we can continue to support them in the meantime while they prepare and make those plans.
And I think the reality is that because of the fact that people are being forced to migrate anyway, it does give them the chance to have a look around and see what’s available in the marketplace and explore alternatives and see if there’s maybe a better option because they’re having to do the migration anyway, so what difference does it make if you migrate to something better versus more of the same?
Heather Brodbeck:
Mm-hmm. Okay, thanks. And related to that, is there thinking about the flexibility of the different platforms. I mean, I myself have had the experience of migrating from one data warehouse to another. Is there anything about the IBM solution that’s more or less flexible than some of these other options that are out there? Maybe Umair, I don’t know. I think you’ve had some of that experience.
Umair Waheed:
Sure. Looking at… I do a lot of work on competitive and comparative analysis. When we look at Netezza’s solution in the cloud, at least how it was first proposed, it’s very much, it is a lift and shift from their on-premise capability. And anyone that’s built a cloud service knows that cloud works very differently to how on-premise works. It’s very difficult just to take an on-premises technology and lift and shift it to the cloud. It looks different, it feels different, it’s managed very differently. And the technical response to when you try and improve performance is very different as well.
I think one of the things that Netezza does in the cloud is that it requires you to run always on. In the cloud, we want flexibility, we want to be able to move things up and down elastically. And the way that Netezza is currently architected with the tight coupled storage and compute doesn’t really work that way.
And we’ve started to see them move away from that slightly in some of the newer releases they’ve done on Azure, for example. But they’re still learning and it’ll take them a while to get there. A great database, great cloud data warehouse doesn’t get built overnight. Even with the best of engineers, it takes time. And we know that. We’ve been building our cloud technology for the last three or four years and it takes a while.
Heather Brodbeck:
Yep, yep. All right. Thanks Umair. Why don’t we switch to Asim now, just on the topic of other options. You’ve been through this, you’re in the middle of it right now, helping the Navy with their migration. Maybe you could talk a little bit about that process. What are some of the features, what’s the testing involved and what alternatives did you support them in looking at?
Asim Aziz:
Yeah, absolutely. I want to touch really briefly on what you and Umair mentioned. We went from one petabyte Netezza to 1.5 petabyte Yellowbrick, and our footprint to Umair’s point is reduced now from four or five racks to less than a rack. I look at it and I say, how is that even possible?
And we have double our performance and ETL workload and still we are better on performance numbers and the rack is not even filled yet, a single rack. That’s just the modernization from modernization aspect. I think that Yellowbrick continue to do innovations in this field, whereas I think Netezza just was so good that they just said, “Okay, it’s good enough for us for right now.”
From a process perspective, it took us… This is one of the most important system that my customer has and a lot of… It’s an important system from production perspective. We serve almost all major branches within the Navy on this system. It was about a six-month process for us to pick an alternative. When we were looking for an alternative, our number one goal was we need to find somebody who has a long-term strategy, not just on premises, but also in cloud. And we didn’t want to do another migration in the cloud, so we were looking for a same core base. Once we do migration on-prem, we can just do a copy paste in the cloud.
And then I think we were also looking for a good partner who has skin in the game, so to speak, who can help us from migration perspective, a technology that is easier. As you mentioned, migration is a migration. I mean, no migration is easy, but it’s just whatever is easier. We have, I mean, a lot of source systems from which we are getting data. We have some systems that we have been using for 18, 20 years on ETL. We need to make sure that, without modernizing the ETL scripts, we can still get the data in. And then when you do migration, it’s a different technology, different appliance.
You have to touch pretty much every aspect of core that you have in your enterprise. To Huw’s point, we can automate things, that’s where we are. Yeah, we are in the mix of it. We’re really excited about it. We have great support from Yellowbrick engineers, so we have had no issues. And if there are any technical issues that came up, they were resolved.
Heather Brodbeck:
Great, thanks, Asim.
Asim Aziz:
Thanks.
Heather Brodbeck:
Maybe Huw, you have the 30 plus years’ experience. If you could maybe just give some insights into migrations that have attempted or failed or ones that have been successful. What are some of the things that were key to those and things to watch out for? And then maybe a little bit around just total cost of ownership for a data warehouse.
Huw Ringer:
Well, I think that one of most difficult aspects of a migration is the validation side of things, the verification, to make sure that you haven’t lost any data in the process of the migration itself, or haven’t corrupted it or truncated values or none of the integrity of the data has been lost in the migration process. I think that’s a key issue. It’s not just a question of just lifting and shifting it and then you’re done. You have to prove that the data is right.
And another complexity really is also the fact that when you have, as Asim said, petabytes of data or even hundreds of terabytes of data, you can’t really practically go for a big bang migration necessarily all in overnight.
Typically, what you have to do is phase the migration over time and run the original system and the new system in parallel side by side while you perform acceptance testing and then cut over time. And so the challenge then becomes how do you keep the two systems in sync while you are planning and while you are performing the cut over between the two? Those are some of the challenges that we have addressed and that we have helped customers with in the past. In terms of… Sorry, what was the second part of your question?
Heather Brodbeck:
Just thinking about cost of ownership, and I think one of the things that Asim was mentioning there that they’re on-prem with the full intent to move to cloud. One of the things I think about is how much more money am I spending to do on-prem first and then go to cloud?
And obviously, there’s the benefit of being with the same vendor, so that transition shouldn’t be one for one for the build on-prem with the migration cloud. But I have to imagine that there is additional cost there. And maybe the question might be just how should companies think about that extended cost of doing the on-prem and then to the cloud with the vendor?
Huw Ringer:
Well, I mean, certainly we have had some customers that have been quite badly burned, I think, by just trying to dash for cloud. We’re talking a couple of years ago now. And in the same way that a lot of customers embraced Hadoop perhaps a little early, and then also found that to be quite a steep learning curve and not as performant as they had hoped and a lot more complicated and fragile and brittle than they had been used to with Netezza. We’ve had other customers that had no conception really of how much it would cost to be metered by the bite. Every time you run a query, every time you download an answer set, you’re getting charged for that. And the irony is the more popular and the more the platform and the more people use it, the more expensive it becomes. And so the cost can be a real shock actually to customers if they’re not prepared for that.
And one of the things that I particularly like about the hybrid approach that Yellowbrick has is that it gives the customer the flexibility and the choice to choose which workloads they want to run where. For example, they can do all of their developments, say for example, in the cloud, but deploy all of their production processes on-prem so they don’t have to pay for all the data transfer fees that may be applicable in the cloud, for example, depending on what platform you’re using.
The other thing I like about Yellowbrick is that they don’t try to resell any cloud service provider’s infrastructure. They charge, it’s a very transparent, I think very fair and very simple pricing scheme because exactly what it’s going to cost you right up front and you’re in control of your own cloud spend through whoever your preferred cloud service provider is, and that’s completely separate from the Yellowbrick charging.
I really like that flexibil – Oh, by the way, also, there’s full database replication built into the product from anywhere to anywhere. That means that you can have multiple instances on different cloud service providers. You can arbitrage between the cost of them, you can move your workload from wherever is the most economic at any given point of time, and there are no penalties for doing this. These are all I think factors which customers I think will find attractive.
Interestingly, the reason why people moved to Netezza in the first place is because of their three proofs. There was the proof of price, proof of performance, and the proof of simplicity, how easy it was to administer and manage and maintain the environment. And I think what Yellowbrick does is they deliver on all those proofs. In fact, over deliver on all of them. They’ve taken them to another level.
And as the other panelists said, unfortunately what IBM have been focused on is trying to move to the power PC architecture, trying to get Netezza to run inside virtual machines so they can put stuff in the cloud as well as on-premises. And as a consequence of that distraction, and it’s much harder for them now to catch up to people like Yellowbrick that have got a bit of a head starter on them in some of those areas. I don’t know what you feel, Asim. Do you agree or what’s your experience?
Asim Aziz:
Yeah, I totally agree on all the aspects that you said. From cost perspective, what I like, this was one of the items for us, as you know, from government customers perspective, we cannot have a variable flexibility on pricing. What we liked is that Yellowbrick doesn’t charge, at least for DOD customers in the cloud, a variable price. You just get your cloud infrastructure and you’re paying however many transaction you want to run, unlike other cloud vendors that are charging by meter. That’s one other deciding factor for us was going to the cloud.
Heather Brodbeck:
Great. All right. Thank you. Let’s go, Huw, back to you. You obviously, CEO of Smart Associates, many experts from Netezza in your organization. What do you feel is potentially one of the best alternatives for Netezza for your customers? And maybe just some information on some different partnerships that you have.
Huw Ringer:
Well, I mean, we really love Netezza as we helped IBM… Sorry, we helped Netezza get set up in Europe originally. We’ve supported every customer on the planet. We used to be part of the global Netezza support team. Pretty much everybody on this call who’s a Netezza customer, we will have supported you at some point in the past.
We are partners with IBM, but like I said, the problem is that a lot of customers in our experience are done with IBM. They feel betrayed, they feel let down. They’re looking for actively for alternatives. And actually, how the Yellowbrick relationship came about is because we had some Netezza customers come to us as, “Hey look, we want to move to Yellowbrick, but we need someone to support our Netezza environment while we do that. Can you please provide the support and help us with the migration?” And that’s how the relationship with Yellowbrick started.
We do play nicely with other vendors as well. We have worked with Snowflake and Redshift in the past. Our data migration tool works with other technologies like Google, Big Table and so forth. We have done quite a bit of work in this space.
And the reason why I’m on this conference call today is because what I’ve seen is for customers that have already migrated to Yellowbrick, the performance improvements that they’ve seen, the cost savings that they’ve seen, not just relative to Netezza, but relative to other cloud database providers, and the simplicity as well of the environment. These are the reasons why I think the Yellowbrick option is very attractive to Netezza customers.
As well as frankly, the broad compatibility because Yellowbrick is PostgreSQL based, just like Netezza is. A lot of the commands and syntax are very, very similar and doesn’t require a huge amount of effort, frankly, to migrate to. Yeah, so I think all different platforms have their own relative strengths, but I don’t think there’s anybody that does what Yellowbrick does in terms of that multi-cloud on-prem hybrid combination of price, performance, simplicity. Over to you.
Heather Brodbeck:
Great. Thank you. I’m glad we have this recorded. That was amazing. Thank you.
Huw Ringer:
It’s true, right? That’s the thing. It’s true.
Heather Brodbeck:
I know. I mean, it’s why I work here because I believe in what we’ve got. I think it’s amazing. Maybe just to dovetail on that a little bit, Umair, I think we were going to share, and I know we’re getting close on time here, but maybe just a few customers that we have helped to migrate off Netezza. I think this for a long time has been a little bit of our sweet spot given some of the stuff that Huw shared about the compatibility there and the simplicity of it.
Umair Waheed:
Yeah, absolutely. Thanks, Heather. Yeah, we’ve seen a number of customers, helped a number of customers migrate entities. Some of the logos are up there on the screen right now, and you’ve got companies serving huge numbers of retail, analytics firms doing deep retail, analytics like Symphony Retail, providing a bureau service. They’re organizations like LexisNexis who are serving applications on a B2B applications on the back of Yellowbrick. And Yellowbrick has opened up some of these product opportunities for them. Some of these companies have been using Netezza since its inception, companies like TEOCO, and when you look at some of their testimonials on our website, they’re talking about doing things that weren’t previously possible. Getting insights and analytics they just couldn’t do before, either in terms of the complexity, they never came back, or in terms of the scale, the number of people in which they can allow to make those queries and do the analytics through the applications that they serve up.
A broad spectrum of customers across different industries, but all with the common, take some of the common threads that Huw mentioned, the simplicity of migration, the simplicity of having something that looks very similar to Netezza and that central bridge to on-prem to cloud.
And actually being able to decommission their previous platform, which is the biggest challenge that I’ve seen in my career as people leap into the cloud, is that they get to the cloud, but they’ve still got their on-premise thing. They’ve not reduced their cost, they’ve doubled or tripled their cost in some cases as the support for that legacy platform goes, gets more and more difficult to maintain.
Huw Ringer:
By the way, can I just say that some of these names on the slide here, these are some of the biggest Netezza customers, some of the earliest Netezza customers as well. This is non-trivial, the fact that you have these names up there, as well as of course the DOD and the Navy.
Umair Waheed:
Absolutely. Yep.
Heather Brodbeck:
Yep. All right. Thank you, Umair. Well, I think we’re about at time, so just to wrap up, special thank you, Asim, to you for joining us. It’s great to hear about your hands-on experience with the US Navy and how you selected Yellowbrick. We did spend probably a little bit more time talking about Yellowbrick than was in the plan, but appreciate the commentary there, Huw, from you. And also, I mean, clearly with your background and experience, I think you’re a great resource for anybody on this call to reach out to if they’re in this situation with Netezza for end of life. I think your guidance would be invaluable to those individuals and how they think about that migration.
And obviously, thanks Umair for joining today and giving your insights also. If anybody has any questions after this session, you could definitely reach out to any of us with those questions. Be happy to answer. And I think that’s a wrap for today. All right.
Asim Aziz:
Thank you.
Heather Brodbeck:
Okay, thanks everyone. All right. Bye.
Huw Ringer:
Lovely to meet you again.
Heather Brodbeck:
You too. Bye.
Huw Ringer:
Bye bye.