How NetApp transformed the IT infrastructure of Revlon

On this podcast we talk to Lawrence James, Product Alliances and Solutions Marketing Manager of NetApp who have been working with the beauty manufacturer Revlon to transform their IT Infrastructure. Revlon are a very large organisation with manufacturing and products in 100 countries across 6 continents. Lawrence will be talking us through some of the things they implemented and how similar organisations can learn lessons from the technology that they implemented.

For more related podcasts click here.

To subscribe to receive new podcast episodes for free click here.

So firstly Lawrence explain the challenges that Revlon were facing with their IT Infrastructure prior to NetApp coming on board?

One of the biggest challenges they have coming from a business perspective is the visibility of the business across multiple geographies. They have got a massive business and a huge number of product lines into the hundreds and millions in terms of the products they have to follow and be able to measure in each of these geographies. The challenge for them was getting an insight of the business from the point of view of knowing where to invest and where not to invest, insight into the product line and insight on the components that make up their products as well. So in a way what we are looking at here is a big data story and what we were aiming at was the analytics engines that they need to put in place to actually improve their view of their business worldwide so it was quite a challenge for them.

And cost reduction must have been a key driver for change here as well?

Yes, significantly, they were aiming at getting costs reduced; just take the energy costs for instance. The energy costs were reduced by over 70%.

What were some of the solutions you provided them with to achieve these savings then?

NetApp is a storage company and our prime focus is in making storage efficient for our customers. This is often a big challenge for lots of organisations and you can imagine that for many customers when you are developing new products to market and developing new business elements you create lots of duplication and the infrastructure is not used as efficiently as it might be. Just take for example what we do in our everyday lives; We create lots of copies of say production databases and on our laptops we create lots of copies and duplicates of files so just doing simple things like de-duplicating data and compressing data reduces the actual physical space the data takes up. So those are some of the key things we focus on but also automatic things such as applications in virtual hosting environments. Analytics of the business is absolutely key to Revlon given the size of market they must monitor. Things like uptime and availability are very important. They were looking for something near to six nines of uptime and I did actually sit down and work it out that it means achieving less than 13 seconds of outage a year, so quite extreme in application uptime. From NetApps point of view one of our key things is to ensure through things like clustering that a customer such as Revlon can achieve such levels of availability as this.

IT as we all know breaks, you know you will get elements that fail especially on the mechanical side where you have got mechanical disc storage. But, these days it is perfectly possible as it is in a compute layer with virtualization and VMware to maintain a connection from the application to its data and get through those failures without stopping the application and certainly that is one of our big focuses.

What has been the impact on their business from these changes?

In large organisations when you have to develop a new business product, bring that to market and increase velocity of that application, the development cycle is key. They have seen an increased projects throughput of something like 425% and that has enabled them to deliver products ahead of time and get ahead of the market and get ahead of their competition, which is often key for a lot of us these days. Decreased time to market means the majority of their projects are now seeing well over 99% improvement of product projects hitting their budgets and being delivered either ahead of time or on time.

Well one of the statistics that caught my eye from this is that you managed to reduce energy costs in the data centres by 72%. That is big money for a company on the scale of Revlon isn’t it?

It is! You can imagine that one the biggest costs within the data centre is from what sucks the energy out of the wall sockets through spinning all these mechanical discs. They are energy hungry and they product a lot of heat so you have to cool them as well and often to get performance and efficiencies with other infrastructures, we don’t use them particularly efficiently. So our goal is to enable customers such as Revlon up to 50% physical capacity on their machine rooms and data centre floors and that is achievable by improving the utilization of the storage they have got just by intelligently what I would describe as high function data management features such as de-duplications, compression and thin provisioning. All of these things combined together can significantly reduce the amount of investment needed and the amount of capacity you put on the floor. Also changing that dynamic are things like flash technologies as well. It is just ramping hugely now. Flash has kind of come of age if you like. When you combine it with traditional mechanical hard disc technology you can use it in different ways it is very powerful. For instance, you can use it for caching, so you cache heavily in the red blocks in the flash layer and you can get some extreme performance out of it. This reduces the requirement for having to put such a physical hard disc on the floor for a lot of our customers. You can reduce the spindle count, the number of hard drives you have to put in the system to achieve a given level of performance, so I see that as a game changer within the industry.

Has the power of these benefits come from combining flash and cloud together?

As the adoption of flash becomes greater it will polarize the storage market. The price of Flash Technologies continues to drop. I had a look at some of the stats around it the other day and in 2008 you were paying something like 65 dollars per flash gigabyte. Today I did a quick search around for an enterprise type of Flash and last week you would be paying something like 8 dollars per gigabyte, so the price reduction in Flash has just been phenomenal over the last 4 years. That is really set to continue and what that will do is eventually polarize the market and more and more workloads will come onto the flash tier as that price drops.

Do you think the days of physical storage are on their way out with these developments?

Not yet I don’t think, I think that there are absolute valid reasons why I would want to keep my “crown jewels.” By that I mean the data and the assets that describe my business. I may or may not want to release those depending on regulation or compliance issues. If these were bound in to a service level agreement with assurances that the company that is going to provide those services that would be fine, but I don’t think we are there yet.

However, I think there are elements of the business that are development elements, you see developers increasingly going to the cloud because they can spin up, spin down, spin up very quickly different development type of environments. That is about time to market but for the production and customer facing environments I think there is still a lot of scrutiny of whether we can do that today.

What do you think are the main barriers that prevent some organisations taking the plunge and res-structuring their IT systems?

One big thing that affects a lot of the customers I deal with is transition. How to manage the lifecycle of an infrastructure, retire old infrastructure and bring in new infrastructure. That entire process is a challenge when cloud down time is a rare commodity. In other words, you go to the business and say I need to take the system out of production for 2 hours on a Sunday night; you cannot even do that anymore. I have heard of organizations that have kits sitting in their warehouses unable to bring it into production because they can’t get the planned maintenance windows to do that and when they do get them they have to meticulously plan the time. But, with our solution that is possible today however and that is what the likes of Revlon are looking for in that 6 to 9 of uptime figure. So, times have changed over the last 30 years and I think that is only going to increase to the point where downtime is just not acceptable at all.

What can organisations learn from the approach you have taken with Revlon here?

I suppose the key message for me and the one thing I have learnt over my many years in this business is that actually storage, when you begin to look at it is pretty much foundational. It does represent the foundations on which you build your business on and if today you have a very siloed infrastructure for your storage and actually your compute and your networking as well then you do need to consider how you can unify that and how you can share it. If you are designing “green field” from the ground up then it is easy to take that approach going forward. But, if you are sitting in a very siloed infrastructure then it is a lot more difficult. However, you can make small steps to actually get towards a shared unified infrastructure and start to introduce clustering but, absolutely you should just take on board that the storage is pretty much foundational and you can drive huge efficiencies if you get that layer correct.

What are the technologies we need to be watching for in the future – how do you see things progressing for big scale business infrastructure?

Flash changes everything for me. It will change the whole dynamic within the storage industry so I would absolutely advise everyone listening to this to keep your eyes on that developing market. It is quite confused and quite exciting at the moment and there is a whole bunch of new technologies that are following hard on the heels of flash so if I was going to say one thing I would say that semi-conductor based storage technologies are the future and that is going to be the one to watch. From the cloud side of things the future lies in how cloud enables the business and it brings IT back into the fold of being an enabler for the business if that makes sense. I think we have gone through this curve where the IT function within the organization has been very much a cost centre. But, IT can absolutely be the enabler for the business and cloud is probably the vehicle for that. The way that I see that happening is with things like self-service, automation of provisioning and service catalogues. Also the measurement and reporting and show back, charge back elements are going to be key in the cloud. Storage plays a huge role in that, how can you provision, how can you automate provisioning, how can you do that against a service catalogue of specific levels of service. Then how can you measure and report on all of that. It is absolutely foundational for cloud once more.