Thursday, February 19, 2015

Kony executive Burley Kawasaki on best tips for attaining speed in enterprise mobile apps delivery

The next BriefingsDirect enterprise mobile strategy discussion comes to you directly from the Kony World 2015 Conference on Feb. 4 in Orlando.

This five-part series of penetrating discussions on the latest in enterprise mobility explores advancements in applications design and deployment technologies across the full spectrum of edge devices and operating environments.

Listen to the podcast. Find it on iTunes. Read a full transcript or download a copy.

For our next interview we welcome Burley Kawasaki, Senior Vice President of Products at Kony, Inc. The discussion is moderated by me, Dana Gardner, Principal Analyst at Interarbor Solutions.

Here are some excerpts:

Gardner: Before we explore the Kony World news, what's going on in the enterprise mobility marketplace? What are enterprises looking for in their mobility strategy?

Kawasaki: Obviously, mobility has proven that it’s not just a passing fad. It's really evolved over the last four or five-plus years. Initially, most companies were just trying to get one or two apps out in the public app store.

Kawasaki
Many started with some type of branded consumer apps, what are called business-to-consumer (B2C) applications, and they were willing to make the investments to make it have a fantastic user experience. They would try to make this a way for customers to experience and engage the brand. A lot of times you saw this being built and launched by the marketing organization inside an enterprise.

Now, what we're seeing is a shift. As people are looking for the next set of ways to exploit mobility, they're looking internal, inside their enterprise. They're looking at what I refer to as or business-to-employee (B2E) applications.

But instead of one or two apps, there are literally dozens or hundreds of mobilized processes and applications that most larger enterprises are looking to build as they start looking at all the internal processes. It could be mobilizing sales or employees, looking at support out in the field with field technicians, or providing self-service access to vacation requests.

There are a number of challenges this creates. One is lack of skills. If you're building one or two, you can probably muster the technical expertise or you can outsource and hire an agency or someone to build it. If you're looking to supply dozens -- some larger enterprises are looking at hundreds of internal-facing mobile apps -- that really highlights the imbalance between the demand from the business stakeholders and the supply of IT skills, resources, and technical talent.

Build applications quickly

Kony, since day one, has focused on how to drive faster and faster acceleration of the full development process. That's part of our core value proposition of rapidly delivering great mobile apps by providing tools and platforms to help build applications more quickly.

When we talk about building anything custom, there is a certain amount of time, typically three to six months that you spend, not just for the development, but to map out the requirements to do all the testing and final deployment. And with any custom software development, you can only compress it so far, and there's a certain amount of skills and expertise that you need.

To answer your question, we think that there needs to be other types of models for ultimately creating these internal mobile applications. The trend that you're starting to see, and that we believe is really going to take off, is a move away from custom, bespoke development of each and every app, to much more of an assembly and configuration model.

If you look at building a home, for example, there was a time where you had to custom build all of the parts to your home. You would go out, cut down the trees, and do everything from scratch, but that was a hugely inefficient process.

Now, essentially, homes are componentized. You can find standard sizes lumber parts. Large parts of your home may be prefabricated and it's just a matter of assembling and configuring them to meet your needs.
Many industries have realized the benefits of moving to assembly and configuration, as opposed to custom built.

We've seen the same assembly across a number of industries, like the auto industry. Many industries have realized the benefits of moving to assembly and configuration, as opposed to custom built.

We're seeing this in software as well. There was a day where everyone used to build their own enterprise resource planning (ERP) system or their own sales automation system. Now, people have moved to the configuration of packaged software. Mobile applications are now at the tipping point where they need to have a different way that will address the explosion in demand that I was describing.

There are a couple of things that we think are required to create this new model. One is that you need to have an ecosystem that provides pre-built components. Obviously, you can't assemble things if there is nothing to assemble from. So there needs to be an ecosystem of components.

Then, there needs to be some type of tooling that allows you to assemble the components without having to be a developer, but more of a visual drag and drop type of composition experience.

And then once you have done that, it can't just be a pretty picture. It needs to actually somehow run and make its way down to your phone or to your device. So there has to be some type of execution or dynamic run capability behind the description of what you have created.

Those are the three requirements. Of course, we have just announced this week some software that addresses each of those categories.

Major announcements

Gardner: Well, let's delve into them a little bit. There were three major announcements around your Marketplace, your Modeler, and also an example of how these come together in your first prepackaged application called the Kony Sales App.

Kawasaki: I'll talk about each of these. I'll start with the Marketplace. As I said, to make this practical and useful for our customers, we need to be able to create a way to find and discover pre-built components. Some of these components Kony may build ourselves, but we're also working with a number of very talented leading edge partners of ours -- independent software vendors (ISVs) and systems integrators (SI’s), who are also contributing prebuilt components.

This week, Feb. 4, we launched Marketplace. If you go out to community.kony.com/marketplace, you can browse. We're adding partners on an ongoing basis, but you'll see some of the early solutions that are available in the Marketplace. That’s the first part of the announcement.

The second piece is around how to assemble these into an actual application, a new product called Kony Modeler. Unlike some of our prior products, our developer tools, these do not require development backgrounds.

The typical profile of a user of Kony Modeler would be either a business analyst or someone closer to the business who knows how to drag and drop, to define what the end-user experience should be for your mobile app, knows how to describe the process or the workflow that has to occur, knows how to take those forms that they've painted, and be able to map it to some backend business data, coming from a system like SAP or Salesforce.
Unlike some of our prior products, our developer tools, these do not require development backgrounds.

As long as you can do that, you don’t have to be a developer and drop into code. You can describe this visually. You can drag and drop. Then, when you're done, the important thing is that it’s not just a picture that you print out and you throw over the wall to your developer. This description of your application then gets pushed out instantaneously to our cloud run time.

We've extended our backend-as-a-service, what we call Kony MobileFabric, so that it takes this model, this description of the mobile app, and will download it to your device and run it. Then, the next time as an end-user, if you are using one of these apps, you just automatically get whatever changes or updates have been made. You don’t have to go out to an app store and find a new app. It just automatically is part of your app.

As an analogy, in the same way if I use any software-as-a-service (SaaS) software, I won't have to install a new app on my laptop. I just go out to my web browser, and next time I log in, it's always up-to-date.

Gardner: It sounds as if this has some of the greater elements of platform-as-a-service (PaaS), but the tooling is designed for that business-analyst level. It also gives you some of those benefits of rapid iterations. You can change and adjust. You can customize to different types of user within the group that you're targeting. And all of this, I assume, is at also low cost, given that it's a SaaS based approach. Tell us a little bit about why this is like PaaS, but PaaS-plus.

Non-developer experience

Kawasaki: PaaS typically has been targeted primarily toward developers. And it’s maybe a higher level productivity for developers, but you still have to write code against software development kits (SDKs) or other application programming interfaces (APIs). Kony Modeler provides a non-developer experience.

The other big thing, and you pointed it out, is that it really does lower all of the infrastructure, hardware, and software costs that are required, because it’s purely cloud-based. It makes it not only lower cost from a total cost of ownership (TCO) standpoint, but it also accelerates the whole development cycle.

I think about this as a shift away from a classic waterfall-type model, to much more of an agile model. In the old model, you spend three to six months trying to go through and nail the requirements and hand it off to your dev team. Then, they go off, and you find out, only when it's in final QA, that it doesn't look right on the device, or it comes back and the business has changed their mind. That never happens, right?
It makes it not only lower cost from a total cost of ownership (TCO) standpoint, but it also accelerates the whole development cycle.

Modeler allows you to very quickly iterate a working application to release in a matter of days and be able to do testing with your end-users. Based on their feedback, I can make updates on an agile basis and continuously iterate on functionality or enhancements to the application.

Gardner: Burley, it also sounds like you're able to bring A/B testing type activities to a different class of user, where you don't always know what your requirements are precisely, but you can throw things on the wall, try them out, see what works, and iterate on that. I don’t recall too much of that capability being available to a business analyst type of user.

Kawasaki: You're correct. Usually, there is this very extended process, where a business analyst has to document everything in some thick specification, and even if you have it wrong or you are uncertain, whatever you communicate out to the dev team is what they go off and build.

So it’s not that this does away with requirements, but it does allow more flexibility to change or to test. And I'd agree. I think the responsiveness will allow much more experimentation and innovation. It's better to fail fast. If you have tried something out and it's not delivering the results, you haven't invested a huge amount of time and cost to learn that.

Gardner: And another appealing aspect of this for IT and operations is that this isn't shadow IT. This is under the auspices of IT. They can bring in governance. They can audit as necessary and make sure the right backend sources are being accessed in the right way, with the right privilege and access controls. They can monitor security. We talked about how it's better than PaaS, but it's also better than shadow IT for a lot of reasons.

Lack of skills

Kawasaki: It is. We were talking earlier about the skills shortage, and if you look at the stats or the data, most industry analysts predict that up to 60 percent or 70 percent or more of mobile development is outsourced today, to either an interactive agency, a systems integrator, or someone else, because of lack of skills.

So it has been outsourced to some third party, and who knows what technologies they are using to build the app. It's outside the typical controls or governance of IT. So it's not only shadow; it's dark matter. You don't even know it exists; it’s completely hidden.

Yet, at some point, inevitably, those apps that you may have outsourced for your first version, it’s not just a first version release. You want to update it sometimes monthly. So it has to come back into IT at some point, for no other reason than it's connecting and talking to enterprise data in the back end. It's connecting to other IT controlled systems, and so there is a huge amount of risk and costs associated if these things are completely hidden off the grid.

Gardner: Let's take this from the abstract to the concrete. We actually have an application now in play called the Kony Sales App. Who is that targeted to, how does it work, and what do you expect to be some of the proof point metrics of this in usage compared to how organizations conduct themselves with customer relationship management (CRM), especially if there is multiple CRMs in play in an organization?

Kawasaki: That's a great point. First of all, this is the first of a series of what we call ready-to-run applications. And the reason we call it ready-to-run is that it's a packaged app. This isn't a custom or bespoke app, but it's pre-connected and pre-integrated to the common back end, in case of CRM what most companies are using, something like Salesforce or SAP on the back end.
So we've taken a task-oriented approach and created a modular micro app approach that really is meant to be very easy and engaging for the end-user.

So it comes ready to run, but like packaged software or SaaS software, it allows you the ability to configure and customize it, because everyone’s sales processes or their user base is going to be different. That's where the Modeler tool allows you to configure it.

So when you purchase Kony Sales, you get not only the application, but the use of Kony Modeler to be able to customize and configure it. And then, as you make changes, you push it live, and again, it deploys using the SaaS model you were describing.

To talk a little bit more about Kony Sales, we think it's a new style of mobile apps, what I will refer to as a micro app. Historically, people thought of CRM software, and I am overgeneralizing, but as big, somewhat monolithic, applications.

One of the historical challenges with CRM usage is that you had to bring your laptop with you, and sales reps are notorious at not completing data in a timely fashion. It takes a lot of mandates, top-down from the sales leadership, to get data into the system so you can get accurate reporting. It's one of the age-old problems.

We believe that if instead of trying to get the whole CRM application crammed down onto a four-inch screen, with all the complexity that it requires, you target very specific action-oriented micro apps that a sales rep can do very quickly on the go, that doesn't take a lot of training, and doesn't take a lot of thought. They can very quickly look up and see their accounts, or they can very quickly log a call they have made.

So we've taken a task-oriented approach and created a modular micro app approach that really is meant to be very easy and engaging for the end-user, which in this case is a sales rep.

User experience

Gardner: And again, for the understanding of how this all works across multiple endpoints, regardless of what your sales force is using for their mobile device, this is going to come down. They are going to get that user experience and that interface that the craftsmen behind the app demanded and designed.

Kawasaki: That's right. Kony Sales is multi-channel. It works across phones, tablets, iOS, Android, and importantly, it does not replace your existing CRM data. It extends the CRM systems you already have, but makes them much, much easier to very quickly get access to.

Also using Mobile First types of approaches, and by that I mean if you are a sales rep, very likely you are on the road or in an airplane. How many people have tried to use whatever CRM client, even some of the web mobile experiences, to get data into Salesforce or SAP? It's all web-based, HTML5-based, and it doesn’t work if you're not online.

One of the things we designed in from day one was that you have to be able to operate in an "occasionally connected" mode. So if you are offline, either because you're out in the field talking to your customer, or you're in an airplane you can still have the same easy access. Then, when you're connected again, it will synchronize and handle updating SAP or Salesforce in the background.

Gardner: Now that we have the model of the Modeler, the Marketplace and these ready-to-run apps, what comes next -- more apps, bigger marketplace, or is there another technology shoe to drop?

Kawasaki: It's more apps certainly, and not just from Kony, but from our partners. When we did some of our initial planning and research, the most commonly mobilized processes were ones that were customer facing or customer impacting, just because of the benefits and the ROI.

So we started with sales. We're going to release our next one, which will be around field service. It really helps engage at the point that you're supporting and serving your customer.
It really helps engage at the point that you're supporting and serving your customer.

There are a set of these that we are working on, but I think also importantly, we're working on really making our partner ecosystem trained, ready to use Modeler, and to build very unique and differentiated applications to publish to the marketplace.

We have a couple of examples of these ready-to-run apps that are compelling from our partners that you will hear more about, and that list will continue to grow over the coming weeks and months.

Listen to the podcast. Find it on iTunes. Read a full transcript or download a copy. 

 You may also be interested in:

Wednesday, February 18, 2015

Mexican ISP Telum gains operational advantages via better monitoring across vast network elements

The next edition of the HP Discover Podcast Series delves into how Telum in Northeast Mexico improves their ISP services delivery reliability through quality assurance and higher availability using advanced monitoring software.

Listen to the podcast. Find it on iTunes. Read a full transcript or download a copy.

To learn more about how they have matured their process, technology and IT culture, we are joined by Max Garza O'Ward, Head of IT Operations at Telum, an ISP based in Monterrey, Mexico. The discussion is moderated by me, Dana Gardner, Principal Analyst at Interarbor Solutions.

Here are some excerpts:

Gardner: Why are reliability and high performance are so important to your business?

O'Ward
O'Ward: It’s a very competitive business for the telecommunications industry. We're not the top dog in Mexico or Northeast Mexico. Everything that we put into our customer effort is to bring in customers that we need to keep. So reliability is a key part of our commitment to our customers.

Gardner: And you're not just using technology. You're delivering technology. So it seems essential to have a handle on what you have, what it's doing, and maybe even get out in front and have predictive capabilities when problems might arise.

O'Ward: That’s very true. Prediction is where we need to focus. To ensure good services we need to make sure that all those systems are up and running, and we use software to precisely do that.

Gardner: For our listeners who might not be familiar with your organization, tell us about your size, how many subscribers, how many services. Just give us a description of your organization, both in terms of the breadth of services and the size of your audience?

O'Ward: We're part of the Northeast Mexico market, basically Monterrey, which is the biggest city up north. We have three different customer markets. The residential market is roughly over 500,000 customers. We have small business or SOHO businesses, with between 3,000 and 6,000 customers. And we also have a large enterprise market, around 1,500 large enterprises.

Unwieldy network

Gardner: Let’s dig a little bit into the problems that you face. Several years ago, you were looking at a network that was perhaps a bit unwieldy, maybe not well-defined. You had some difficulties predicting how certain things that you did on your network would impact your customers. Perhaps you can walk us through your problem set, your challenges, and then how you started to solve them.

O'Ward: That’s a very good approach. In terms of network, we basically started noticing that we were committing a lot of unplanned outages, or unplanned downtimes. So we started to reinforce our monitoring solutions based on HP software. We came in and provided better solutions to what we were looking into as a network element point of view.

Based on that, we refurbished our inventory and made sure that all of our network elements were replaced promptly, based on events. So prediction was key to our better service-level agreement (SLA) offerings.

Gardner: Max, was this a function of changing just the software or was there a cultural component to this? Did you have to change the way you were thinking about monitoring and quality assurance in addition to employing some new technology?

O'Ward: Yes, it was a cultural change. As a matter of fact, just two years ago, we revamped the way the operations department is composed. So a big gap that was closed because of culture. Culture needed to be changed.
What secrets will your data tell you?
In the HP Toolkit for Operations Analytics
Previously, we had all these disparate teams working on only their solution. Once we came under one head of operations, we decided that service was the only thing that matters. So we bridged that gap and now we have all these cross-functional teams working for the same response, which is service offerings.
Now, it's a combination of culture, teamwork, and understanding where the failures are.

Gardner: So IT service management (ITSM) has led to the ability to maintain your quality and performance. Are there any indicators of how much -- perhaps the number of failures from one period to more recent failures?

O'Ward: There are a lot of numbers. I will give you top figures. IT is the department that I head, and most of these departments are based on different engineering groups.

When we started working toward service and focusing only on services, on video services, for example, we had over 10 percent failures globally, not every month, but throughout the year. Once we got under this new management and using our new HP tools, we have been bringing that number down consistently.

Now, it's a combination of culture, teamwork, and understanding where the failures are. Sometimes software tells us where the problem is and sometimes software is needed to understand where the problem is.

In this particular case, we soon understood what the problem was and we decided to change out equipment that was failing via either obsolescence or just a defective part.

Transparency and visibility

Gardner: In addition to changing culture, putting in some better processes and better tools, it seems to me that for a lot of companies that I speak to, a lot of the process involves getting to know yourself better, providing transparency and visibility.

Then, it's dashboarding that information so that people can access it, regardless of whether it’s firefighting or just ongoing maintenance. Tell me about this journey from having a lot of elements, perhaps not always visible, to getting this new-found ability to have greater inventory control.

O'Ward: To start off, transparency is key. Once you have an approach of letting the upper management know where your failures are, that creates concern. And in order for us to create business, we need to have a reputation to uphold.

We started with monitoring basic monitoring elements. We created awareness of where our failures were, and at the same time, we asked for more budget to focus on all these defective parts.
What secrets will your data tell you?
In the HP Toolkit for Operations Analytics
That, in turn, made management very aware of what the engineering departments were actually doing -- either as an IT department or as engineering by itself, which is basically hardware.

Once we had all these components, and they were publicly scrutinized or they were publicly shown in a quarterly meeting, that helped create a dashboard. Now, dashboards are really fun if you really know what you're talking about, but if you give upper management the wrong information, wrong decisions are going to be made. So that’s key.

We're working on creating a huge dashboard. Maybe this year is going to be the year. We have the elements and we're providing that information for the dashboard to be built, but we are waiting to do the next step.

Right now, we're focused on getting the elements straightened out, monitoring all of our key systems, and we have done just that in the last year. So we've upheld our end of the bargain, which is service, quality, and capacity. The next step is going to be providing automatic dashboards. Right now, dashboards are manual.

Gardner: So the good news is that you're getting much more reliable information about what's going on. The bad news is, is you have got a whole lot of information coming in.

O'Ward: That's correct.

Aligning data

Gardner: Big data is a big topic here at HP Discover. What are your thoughts about how to get that data, be it structured or unstructured, into an alignment so that you can improve on your situation, know more about it, get better predictions, and better analysis? I suppose the capstone for this is how important will big data become for you to maintain and improve on your reliability over time.

O'Ward: Big data is a big name, it’s a big trend, and everybody is talking about it. A lot of people, especially people who aren't technology-oriented, talk about it as if if they know it. The way big data is coming into our shop is focused more on customers.

If we're talking about big data, unstructured data, that’s coming in from our traps or alerts and stuff like that. Yes, we need to go into that particular scenario. We're looking at two different projects.

We're going to look into a big-data project that actually brings capacity and quality for our services. At the same time, there's going to be another effort from big data that is a customer-facing effort. So yes, it’s going to be a reality in the next year.
The software is awesome, just great software, and if you have the right people and the right potential, that software can bring you very good benefits.

Gardner: So it’s safe to say that big data is going to have an impact on your IT operations, but perhaps also in your marketing, to understand what’s going on in the field very quickly and then be able to react to it. Big data sort of ties together business and technology.

O'Ward: That's correct. That’s the way we're looking at it. As I said, there are two different teams of people working on it. We're going to be working on the operations part first and then at the marketing part as well.

Gardner: We're here at the beginning of HP Discover. Are there any things in particular that you're going to be out there looking for in terms of how to accomplish your goals over the next several years. What would you like to see HP doing?

O’Ward: Very much what they have been doing in the past. The software is awesome, just great software, and if you have the right people and the right potential, that software can bring you very good benefits.

Our head of operations for the whole company is here with us this week. I'm going to make sure he attends all these meeting in which we can talk about big data and how we can mold out all of the strengths and all of the key performance indicators (KPIs) that he needs. I hope that HP continues to be an innovative software company. I have really enjoyed working with them for the last five or six years.
What secrets will your data tell you?
In the HP Toolkit for Operations Analytics
Gardner: Okay, last question. Going from a failure rate of 5-10 percent down to less than 1 percent is enviable. A lot of people want to make those kinds of strides. Now that you have had experience in doing this, do you have any 20/20 hindsight? What would you suggest to other organizations that are also trying to get a better handle on their systems and their network, get to know their inventory, and gain visibility? What have you learned that you might share now that you have been through it?
O'Ward: It doesn't matter how much we monitor things or how many green lights or red lights we see on any given dashboard. If we're not focused on business processes and business outcomes, this isn't going to work.

My take would be to focus on a business process that you actually know it's critical and start from that. Go top-down from that. That would be the best approach. It's worked for us. It actually bridges a gap between management and the engineering departments. It also provided us with sound budgeting information. Once you understand what the problem really is, it gets approved easier.

So look at business processes first, get to know your business outcomes, and work on that toward your infrastructure.

Listen to the podcast. Find it on iTunes. Read a full transcript or download a copy. Sponsor: HP.

You may also be interested in:

Tuesday, February 17, 2015

California Natural Resources Agency gains agility through Software-Defined Data Center strategy

The next BriefingsDirect innovator case study interview explores how the California Natural Resources Agency (CNRA) in Sacramento gains agility from a software-defined data center (SDDC) strategy.

Learn how this large and diverse agency of state agencies uses the best of centralized and de-centralized IT resources, largely built on a common VMware-powered SDDC strategy.

Listen to the podcast. Find it on iTunes. Read a full transcript or download a copy.

 At the recent VMworld 2014 Conference in San Francisco, our moderator, Dana Gardner, Principal Analyst at Interarbor Solutions, interviewed Tony Morshed, Chief Technology Officer for the California Resources Data Center in the California Department of Water Resources, and Michael Hom, Data Center Chief in the IT Infrastructure Services Branch for the California Department of Water Resources.

Here are some excerpts:

Gardner: You’re from the largest US state. You have a large government, probably the size of what many countries have. Is there something different about the public sector in terms of accountability or response? How do you have a different set of requirements as a public organization and perhaps enterprises, too?

Morshed: We do have a difference. Our procurement is much harder. Getting people is much harder. We live within a lot of constraints that the private sector doesn’t realize. We have a hard time adjusting our work levels. Can we get more people now? No. It takes forever to get more people, if you can ever get them.
We live within a lot of constraints that the private sector doesn’t realize.

Gardner: So it’s doing more with less over and over again?

Morshed: Constantly doing more with less. Part of this virtualization is survivability. We would never be able to survive or give our business the tools they need to do their business without this. We would just be a sinking ship.

Gardner: So the whole philosophy, Michael, of SDDC and virtualization, is of doing more with less, of automating, of boiling out the manual processes and going to more real-time responsive technology driven infrastructure makes total sense for your organization?

Hom: Definitely. To go with what Tony says, we really don’t have much overhead when we need to respond to future projects. When there's an uptick in activity, there is no way to have more resources available. So we need to build that into our infrastructure to allow for that dynamic bandwidth to happen from a personal level.

 Gardner: Help us understand the size of your organization. This is a large state government department, but you're really a department of departments.


http://www.cold.ca.gov/agency_display.asp?ATRID=WTRRSC
Morshed
Morshed: Our department, Water Resources, consists of 3,500 people that are part of the agency that comprises many departments. The bigger ones are Parks and Rec, Cal Fire, and Fish and Wildlife. There are about 28 agencies and conservancies and 25,000 people.

Gardner: So in order to support all these people and all these different agencies, a common infrastructure is important, but also it sounds like you need to have differentiation and customization for their specific needs. How do you accomplish both the goal of a common infrastructure for efficiency, but also still be able to meet all your requirements for all those different people?

Morshed: When we started our consolidation effort, we decided to transform ourselves more to an ISP-style setup. We know that most departments have their own IT shops, and you know there is still that trust thing. So we just built the common infrastructure. We let them share the infrastructure, but they have their own security posture. We segregate all their traffic, so each department can still feel like they're autonomous, but yet we all share the infrastructure, in which case we all share the savings.

Mandate to consolidate

Hom: With that, we had a mandate to also consolidate. First, the State of California is really about cost savings. Each of the 25-plus organizations mostly had their own IT shops. By combining the infrastructure as a service (IaaS) in our multitenancy data center, they're able to reap the benefits of cost savings for infrastructure, but also concentrate each department's specific needs and applications.

Gardner: What are some of the challenges you have in terms of getting closer to a full SDDC? Are these technology, culture, process, or all of the above?

Morshed: At first it was technology, but the culture and the organizational mindset are the bigger challenge. You can find solutions and work through technology. We're IT people, and we’re used to the technology problems. For me, the greater organizational problems -- and the structure of your processes -- become the harder things, because we’re not as used to dealing with those things.

Gardner: Now, Michael, part of the adoption of a complex, long-term journey on IT is to show results early and get buy-in. Are there instances where you look at that? Maybe, it's the DR, where you can have a sense of better dependability on your resources? Any way to describe what you’ve done early on that has lead to a greater emphasis to adopt more aggressively?

Better service levels


Hom
Hom: Definitely. One of the key things with early wins is providing a better service level for provisioning, and that’s something that everybody has been struggling with. With the cloud infrastructure, we've been able to provision within days, if not a day. And that typically beats most of the service tools that each of the organizations have had. So that was an early win.

It’s things like that where we decrease overhead and make IT more accessible for business. That makes it a win, and starts to have the ball rolling as far as other features, such as DR, greater capacity, and things that would be tough for each individual organization to do on their own.

Gardner: Are you exploring software-defined storage, software-defined networking?

Morshed: When the software-defined stuff came out, for me, one of the big things was disaster recovery (DR). If I could stretch my data center into another facility, DR becomes a non-issue, because those workloads can shift between sites without any trouble with automation.

That was the next piece for us -- automation. We realize that we’re part-way there, but to get all the way there, we need to do fuller automation. This means that we need to quit tinkering with the network and storage every time we want to do something new.

Those were big driving factors for going to SDDC. To us, it means that we’re obfuscating the hardware. The hardware’s there. It’s just running. We’re working and tweaking everything at the software layer -- so we could be a lot more agile.

Separating the physical

Hom: For the SDDC we really wanted to provide a logical data center to each of our organizations. We wanted to separate the physical, which allows our folks to support more of a logical infrastructure, where they still have autonomy, but the physical layer is basically one and the same.

Today, from a functional point of view, they get what they’ve had before, but without the overhead of physical support. We've used the VMware vSphere and vCloud Suite to provide that software-defined queuing. Right now, we're embarking on a software-defined networking, using VMware NSX and third-party vendors to support that.

We’re looking to use automation soon to help us decrease overhead, and pass those savings on to each of the organizations.

Gardner: We’re here of course at VMworld 2014. There’s lot of news going on. Anything in particular piquing your interest, perhaps with the OpenStack support in the EVO Hyper-Converged Infrastructure? What is now on your agenda after hearing some news to reach those goals as you describe them?

EVO looks pretty nice

Morshed: There are a couple of things. EVO looks pretty nice. I was out on the floor and looking at it yesterday and talking with the CIO and I see it as something that we might be able to use for some of our outlying offices, where we have around 100 to 150 people. We can drop something like that in, put virtual desktop infrastructure (VDI) on it, and deliver VDI services to them locally, so they don't have to worry about that traffic going over the wide area network (WAN).

The other piece is the acquisition made recently with CloudVolumes and looking at how we can use that to leverage our VDI structure. We're using another product right now in that space, but again with CloudVolumes it’s been a part of VMworld. It’s more interesting, because we know that the chances of all the software being upgraded and updated in at the same time in interoperability is greater if it’s a VMware product.

For us, it’s been a real struggle to make sure that all the products that we use, interact and as there’s an upgrade, everything upgrades at the same time. To me, those are the two biggest things that I'm getting out of the announcements.
The business could come up with more dollars, but to be able to be more agile and more flexible is where it really pays off.

Gardner: Do you have any metrics of success that you can point to from your SDDC effort so far?

Morshed: We do have some tangible benefits. We have reduced our CAPEX by somewhere around 40 percent and our OPEX around 32 percent. I don’t have the numbers, but we have deployed VDI in the Department of Water Resources and we already virtualized about 600 to 800 desktops. Not only is it helping us save costs there; it’s also used as a strategy for a remote access as a strategy to help protect our server infrastructure by using VDI for admins.

So there are those tangible things that you can reach out and measure and those intangible things, where it’s allowing us to do something easier and more flexible. That, for me, is the bigger win. The business could come up with more dollars, but to be able to be more agile and more flexible is where it really pays off.

Gardner: So we get productivity, we have better DR, which reduces your risk, and we've got some hard savings and economics. It's pretty compelling. Michael, any thoughts about how those fit together, and which ones are more important to you?

More flexible

Hom: Definitely, this allows us to be more flexible, and there are some things that we're trying to do that we would never imagine without a SDDC. So they increased security, greater capacity, capabilities to our business.

Gardner: How about VMware specifically? Is there some differentiator in terms of how they produce this products that has allowed you to follow this journey? Is this more of a partnership than a procurement relationship? It sounds like the track that VMware takes in its strategy very much aligns with yours.

Morshed: It’s very much a partnership. In fact, we basically only want to work with business partners. We don’t want to work with vendors, because we don’t need someone to sell us something and walk away. VMware has been hand-in-hand with us for this whole journey.

When we look at other products for the mix, we look for the deep partners with VMware because we know virtual machines (VMs) are core. So, when we look in the storage partners and we’re looking at networking partner, we’re making sure that those partners are partners of VMware.
We don’t want to work with vendors, because we don’t need someone to sell us something and walk away. VMware has been hand-in-hand with us for this whole journey.

One of the things that we find is the inoperability once everything has been virtualized. Everything has to connect, and it’s not a single stack. So if one thing gets upgraded, we need to make sure that everything across a stack can accept that upgrade. Otherwise, we lose the ability to take the advantage of the upgrade until everybody else catches up.

Early on, we were in that position and we’re doing everything we can to remove ourselves from that position.

Hom: We consider VMware a strategic partner. A couple of things that illustrate that is that we’ve been involved with the VMware’s Excellware and Velocity program and that's been two-fold. For the velocity side, we have marked up a fully working SDDC with SX, with virtualized automation, operations in business as a stack.

Gardner: One of the things we've heard here at VMworld is be bold, be brave, be a little bit aggressive. Go out there and do these things. Any thoughts for other organizations that are just dipping their toes into the water? Is it higher risk than reward to be bold and brave in getting early? Or is it perhaps something that allows you to then be a differentiator and be better in your own environment, whether it’s a public or private sector?

Set in stone

Morshed: The first thing is always question what you’ve got that's set in stone, because most of it is not set in stone. We've all heard a lot of things that you can't do. You can’t virtualize Oracle, but you can. You can't do this, you can't do that, you can't get the network focused at the top of the storage. That's all that stuff that you actually can do.

You have to really look at it, peel it back, make sure that "you can't" is an actual thing, and then figure out how to get around it. The way I see it is that, as the world turns, things morph, and if you don’t move into this virtualization space, you're going to be left behind. You're going to be the guy making buggy whips. There are no buggy whips running around. There’s no use for them.

We’re all being asked to do so much more with the same resources or fewer resources. We're all being pushed to keep up with how the demand is going out there. Technology is just jumping, and this is the only way on the infrastructure side to keep up with that.
We’re all being asked to do so much more with the same resources or fewer resources. We're all being pushed to keep up with how the demand is going out there.

Gardner: Michael, any other thoughts in terms of 20/20 hindsight on your experience and why being aggressive and being bold has paid off?

Hom: Virtualization is definitely up and running, at least in state organizations. It’s probably something that we might do that or we might use as a toolset, but from looking at VMware this week, virtualization is the industry standard.

If you don’t take it on, then you really won’t be able to respond to business needs. What happens is that when the official IT organization becomes obsolete, there are going to be ad-hoc IT organizations and those would become the norm. If you want to be relevant, you have to use every tool set that you can to provide the business needs.

Listen to the podcast. Find it on iTunes. Download the transcript. Sponsor: VMware.

You may also be interested in: