Back

Interview with Kent Graziano – on Snowflake, Covid-19 and the Future of Data, part 2

30.04.2020

This is part 2 of our interview with Kent Graziano, Chief Technical Evangelist for Snowflake. If you missed the first part, you can find it here.

Snowflake has near zero administration. However, you still have so many partners around the world. How would you describe their role and the qualities you search for when deciding who to work with?

Importance is in helping our customers who are looking to move to the cloud. They need the guidance, best practices and help during their projects. Many projects are quite large, and they take time. We do have a professional service team, but their job is manly about teaching best practices and doing a bit of consulting. We don’t want to have a huge consulting organization – we’ve always relied on our consulting partners to really help the customers go the final mile. And Snowflake does a lot of things but for each project you still need to think a bit differently, you still need to organize the security from a role perspective, the overall architecture – how do you design within Snowflake ecosystem, what are the best ways to organize your databases, schemata and even doing the DevOps aspect of it. Going from development to test to production. Many of our customers need that kind of help so we really do rely on our partners to provide that expertise.

To a large extent we look for experience of organization and how much experience they have in the cloud – that’s key. Like you’ve talked about your tool for moving data from on prem to the cloud, which is a huge issue. More traditional organizations have no idea how to move to the cloud. Our consulting partners have to be the experts of that, as well all the other things that they know – moving data pipelines, ETL vs ELT, data modeling, understanding data lakes and of course the cloud ecosystem – to be experienced in at least one of the three; either AWS, Azure or Google. If you’re going to help to move people in the cloud, you need to have some expertise in those cloud environments, so you understand how that works.

And then, of course, a successful track record and with many of our newer partners we are looking for people who are maybe even engaged with a Snowflake customer and are familiar with the product or have already sent their people to training courses. Since Snowflake has become very popular and everybody wants to do it, we are very careful because we don’t want to take partners that are not qualified, because that doesn’t do our customers a service. Our customers are our priority and our partners are as well. Our alliance team is working thoroughly on which ones to bring in and who not.

Also, we are growing so fast, so it’s impossible that we alone could hold the hand of every customer.

As you’ve already mentioned, one of the issues for the next few years will surely be how to get from the on-prem world to the cloud. New solutions are emerging to help the process. What, in your opinion, should a good data warehouse automation solution possess?

It really depends on who is using it. I like tools where I can design and really do the architecture in the tool, the high-level mappings and then being able to push some buttons that will generate the code, following a pattern. The tool needs to have a good UI, so it’s also easy for an analyst to use and not just the programmer. It should be also producing documentation, so that you have the data lineage. If you’re going to be compliant with things like GDPR, you need to know where the data came from, where it went and what happened along the way.

It needs to give you the possibility to do things more quickly and iteratively. It’s way easier to change specifications and generate code than it is to edit it. And when someone is hand-editing code, there’s the possibility of human error and a good automation tool should not be generating errors. These kinds of tools allow you to be more agile, to produce better code and have less bugs.

You can design something wrong, so it doesn’t do the right thing but at least it will do what you told it to do. And that’s a little easier to solve than finding something that just doesn’t execute when you have around three thousand lines of code and somewhere in there is a mistake that you can’t find.

So overall, a good tool would allow you to be more agile, to have a more agile life-cycle, to iteratively go through and build your overall architecture and platform a little bit at the time, so that you could be more effective at delivering things faster.

I’ve read quite a bit about you and the feeling I got is that you are a humanitarian by heart. From that point of view, what does the power of data mean to you?

Right now, we’re seeing a great example of it with COVID-19 data. With our partner, company Starschema, who gathered all the raw data and transformed it into something that is more usable for analytics in data science, we put it up in the Snowflake Data Marketplace. To me, that is the ultimate use of data – they call it data for good. I see that there’s a huge potential in there and we’re seeing it right now and it’s very timely with what’s happening. We have had several hundred organizations contacting us to get access to that data.

Last weekend, one of our partners organized a Pandemic response hackathon using the data and a bunch of developers and healthcare organizations participated. A lot of things are happening in that area.

And you’re right, I am a humanitarian and an environmentalist. Both are important – health of the environment as well as health of people. I think there’s just a huge amount of power in having the right data and having it more easily accessible, so that the people can take advantage of it. Pretty much every organization produces some kind of data that would be useful to another organization. So, being able to make that available like Starschema did and putting it out there for other people to access, it’s just a great opportunity to do that.

Snowflake Cloud Data Platform in particular really allows this to happen with the secure data sharing features. When we added that feature a couple of years ago, it was really exciting. It again eliminated some traditional issues. Before, we would need to export the flat files, put it somewhere, sometimes people were emailing it to other people if it was a small data set. Regardless of how that happened, once you’ve downloaded it, you needed to figure out how to get it into your system so you could write queries against it. And even after you’ve loaded it into your system, it could easily happen that something was not right – either the structure, data was missing, etc. To do all that takes a lot of time.

And with things like what we’re doing with COVID-19, every minute is important. Every minute counts because of the fast spread and because of all unknown factors. The ability to get clean, refreshed data much more quickly into the hands of healthcare organizations, decision makers and leaders so they can do the right thing with the most up to date information possible. I think that’s a big win for everybody.

With all the changes rapidly happening in the world, how do you see the role of analytics and data management in the future?

I think one of the things we’re doing now is the model or pattern of what can be done in the future. I have always believed in leading by example and I think this, what we’re doing, is an example. This is COVID-19. If you think about all the other diseases that are out there and other issues in less developed countries. Yes, what’s happening now is big and is affecting all of us, but there are other things out there like that as well. Hopefully, this is setting an example and a pattern for healthcare organizations and governmental entities and governmental leaders as to what else could be done in some of these other areas. Helping undeveloped countries, poverty-stricken areas, helping anywhere there is a disease that needs to be treated at a larger scale. What you can do with data just gives you many more opportunities to respond better and faster and to be more informed.

We know you as the Data Warrior. You have been practicing and teaching traditional Chang Hun (non-Olympic) Tae Kwon Do for almost 40 years. You’ve mentioned on your blog that the martial arts have influenced your approach to your work in IT. Could you explain how?

Yes, my 40th anniversary is coming up in a couple of months. It really just comes from the discipline of traditional martial arts and the ability to be focused. That allowed me to do what I’ve done. As a consultant I’ve heard people say that they’ve never seen anybody as focused as I was. For example, when there’s a problem, I can just sit there and not move for hours, working solely on that problem. In part – and this goes way back into origins of martial arts – is that the physical conditioning and discipline of the martial art is key. To be able to sit and do the kind of work that we do requires a strong mind as well a strong body. To be able to work as intently and intensely, you need to be physically fit as well as mentally focused. You need the energy to do it and get the job done. It’s also my approach to solving problems – continuing, no matter what. And working relentlessly, no matter what that problem is. That’s also the mental aspect of martial arts. It’s called the warrior mentality – if you are in a battle, you need to be focused, you can’t be distracted during the battle. It influenced my entire career, which has been pretty much in IT and quite in parallel with my martial arts training.

Kent, thank you for your time. The offer still stands – if you’re ever in Slovenia let us know and we’ll be happy to take you around !

Ana Mikoš, Marketing assistant @ In516ht

For more inspiration follow the links:

Learn more about Snowflake @ https://www.snowflake.com/

Follow Kent Graziano on his Data Warrior blog @ https://kentgraziano.com/

Or stay with us and learn how In516ht team can help you jump on a cloud: https://www.in516ht.com/what-we-do/services/cloud-first-strategy/

Data Merlin – Data Warehouse Automation Platform: https://www.dmerlin.com/