Gardner: I’d like to hear more about how you put your cloud  together. You're  supporting these services both inside your cloud as  well as your  clients'. Why have you done it this  way?

Rubinow: It’s the convergence of a couple of trends and also things that our  customer started to   tell us. Like a lot of companies, we started to  use cloud technology   within our own company to service  our own internal needs for the reasons   that many people do -- lower  cost, more flexibility, more rapid spin   up, those kinds of things, and  we found, of course, that was very useful   to us.

At the same time, we've talked to a lot of our customers via our commercial division, which we call NYSE Technologies.    By virtue of all the turbulence that's happened in the world,    especially in the financial markets in the last couple of years, a lot    of our customers -- big ones, small ones, banks, brokerages, and    everyone in between -- said the infrastructure that we traditionally    have supported within our own companies, is a new model that we could    adapt, given these technologies that are available, and given that we    NYSE Technologies wants to provide these services. We asked if we should    take a different look at what we are doing and see if we should  pursue   some of these things.

What it comes down right down to  is that   many of these companies said that maintaining their own  infrastructure   is not a competitive advantage for them. It’s really a  cost of doing   business like telephones and office furniture. It would  be better if   someone else helped them with it, maybe not 100 percent,  but like we   propose to do, and everyone wins. They get lower cost and  they get to   offload a burden that wasn’t particularly strategic to  them.

We   say we can do it with good service and at a good  price, and everybody   comes away a winner. So we launched this program  this summer, with one   offering called Compute on Demand, which has a number of attributes that make it different than your run-of-the-mill public cloud.

Higher Requirements

In  the capital markets community, we have some attributes of  infrastructure, a higher requirement,   that most companies wouldn’t  care so much about, but in our industry   they are very, very critical.  We have a higher level of security than an average company would probably pay attention to.

And    reliability, as you can imagine. The markets need to be up all the   time  when they are supposed to be open. A few seconds makes a big    difference. So we want to make sure that we pay extra attention to    reliability.

Another thing is performance. Our industry is very    performance-sensitive. Many of the executions are measured in    micro-seconds. Any customer of ours, including ourselves, are sensitive    to make sure that any infrastructure that we would depend on has the    ability to make sure that transactions happen. You don’t find that in    the run-of-the-mill public cloud because there just isn’t a need for the average company to do that.

For    that reason, we thought our private offering, our community cloud,  was  a  good idea. By the way, our customers seem to be nodding their  heads a   lot to the idea as well.

Gardner: Why have it as a hybrid model?

We're    a very rich source of market data, as one might imagine. We  generate  a   lot of market data ourselves because of the large marketplace  we  are.

Rubinow: In the spirit of  trying to accommodate all the needs that people will   have, for many  of the cloud services, you get the most leverage out of   them, if you  as a customer are situated in the data center with us.

Many    customers choose to do that for the simple reason of speed-of-light    issues. The longer the network is between Point A and Point B, the    longer it takes a message to get across it. In an industry where latency is so important, people want to minimize that distance, and so they    co-locate there. Then, they have high-speed access to everything that's    available in the data center.

Of course, customers outside the    data center certainly can have access to those services as well. We  have   a dedicated network that we call SFTI, Secure Financial Transaction Infrastructure.    That was designed to support high speed, high reliability, and high    resiliency, things that you would expect from a prominent financial    services network. Our customers come to our data centers over that    network, and they can avail themselves of the services that we have    there too.

Historical data

We    have historical data that lot of our customers would like to take a    look at and analyze, rather than having to store the data themselves.  We   have it all here for them. We have applications like risk  management   and other services that we intend to offer in the future  that customers   would be hard-pressed to find somewhere else, or if  they could find it   somewhere else, they probably won't find it in as  efficient a manner.  So  it makes sense for them to come to us to take a  look at it and see  how  they can take advantage of it here.

Gardner: Tell us about your   organization, your global nature, and where you expect to deliver these   cloud services over time.

Rubinow: The full name of the   company is NYSE Euronext, and that reflects the  fact that we are a   collection of markets not only in the United States  but also in Europe.   We operate a number of cash and derivative  exchanges in Europe as well.   So we talk about the whole family being  part of NYSE Euronext.

We   segment our business into three  segments. There is the cash business,   which is global. There is the  derivatives business, which is global,  and  those are the things that  people would have normally associated our   company with, because the  thing we've been doing for many years.

The   newest piece of our  business is the piece that I've referred to  earlier  and that's our  commercial technology business, which we call  NYSE  Technologies.  Through that segment of the business, we offer all  these  services,  whether it be software products we might develop that  our  customers  take advantage of or services as we've already  referenced.

Over    the years, we've been offering these services to our customers, and     then a couple of years ago we decided to do it in a much bigger way,     because we realized the need was there.

In a small way,   over  the years, we've been offering these services to our customers,   and  then a couple of years ago we decided to do it in a much bigger   way,  because we realized the need was there. Our customers told us that   they  would take advantage of these services. So we made a bigger   effort in  that regard. Right now, the commercial part of our business   is several  hundred million dollars a year in terms of revenue.

Question of latency

I have to add one note in terms of latency. For people who aren't    familiar with our obsession with latency, the true textbook cloud    profile means that one could execute cloud-like services. If we had 20    data centers across the world, they could be executed across any of    those data centers and transparent to the customer as long as they get    done.

In ours latency-sensitive world, we are a little bit    constrained with some of the services that we offer. We can't afford to    be moving things around from data center to data center, because those    network differences, when you're measuring things in micro-seconds,  are   very noticeable to our customers. So some of our services could be    distributed across the world, but some of our services are very tied  to a   physical location to make sure we get the maximum performance.

To add further to that, one of the cornerstone technologies, as we all know, of cloud computing is virtualization. That gives you a lot of flexibility to make sure that you get maximum utilization of your compute resources.

Some    of the services we offer can't use virtualization. They have to be   tied  to a physical device. It doesn't mean that we can't use a lot of   other  offerings that VMware provides to help manage that process, but some are tied to physical    devices, because virtualization in some cases introduces an overhead.    Again, when you're measuring in micro-seconds, it's noticeable. Many    other of our services where virtualization is key to what we do to offer    the flexibility in cost to our customers.

So we have kind of a    mixed bag of unique provisioning that's designed for the low-latency    portion of our business, and then more general cloud technologies that    we use for everything else in our business. You put the two of them    together and we have a unique offering that no one else that we know of    in the world offers, because we think we're the first, it’s not among    the first, to do this.

You put the two of them together and we have a unique offering that no one else that we know of in the world offers.

Gardner: So this is a rather big business undertaking for  you. This  cloud is really an instrument for your business in a major  way.

Rubinow: That's right. Sometimes we think the core  of our business is trading.   That is the core. That's our legacy That's  the core of what we do.  It's a  very important source of our business,  and it generates a lot  of the  things that we've been talking about.  Without our core business  we  wouldn’t have the market data to offer to  our customers in a  variety of  formats.

The technologies that we  used to make sure  that we were  the leader in the marketplace in terms  of trading  technology and all the  infrastructure to support that,  that's also  what we're offering our  customers. What we're trying to do  is cover  all the bases in the capital  markets community, and not only  trading  services, which of course is  the center of what we do and it's  core to  everything that we do.

All  the things that surround that  our  customers can use to support their  traditional trading activities  and  then other things that they didn't  used to look to us to do. These  are  things like extensive calculations  that they would not have asked  the  NYSE to do, but today they do it,  because we provide the   infrastructure there for them.

Gardner: What are some of the underlying numbers   perhaps of how this works economically?

Rubinow: From a   metrics standpoint, it's probably too early to provide  metrics, but I   can tell you, qualitatively speaking, the few customers  that we have   that were early adopters are happy to get on stage with  us and give   great testimonials about their experience so far. So  that’s a really   good leading indicator.

Again, without offering  numbers, our   pipeline of people wanting these services globally has  been filling very   nicely. So we know we've hit a responsive chord. We  expect that we  will  fulfill the promises that we’re offering and that  our customers  will be  happy. It’s too early, though, to say, "Here's  three case  studies that  show, our customers are saying how it’s gone,  because they  haven’t been  in it long enough to deliver those metrics.

Many    of the things needed to be done from scratch, because we didn’t  have    models to look for that we could copy in a marketplace.

When  we were putting together our cloud architecture and thinking  about   the special needs that we had -- and I keep on saying it’s not    run-of-the-mill cloud architecture -- we we’re trying to make sure that    we did it in a way that would give us the flexibility, facilities, and    cost that we needed. Many of the things needed to be done from  scratch,   because we didn’t have models to look for that we could copy  in a   marketplace.

And we also realized that we couldn’t do it    ourselves; we have a lot of smart people here, but we don’t have all  the   smart people we need. So we had to turn to vendors. We were  talking to   everyone that had a cloud solution. Lots of vendors have  lots of   solutions. Some are robust, and some are not so robust.

When  it   came down to it, there were only a couple of vendors that we felt  were   smart enough, able enough, and real enough to deliver the things  to us   that we felt we needed to get started. I'm sure we will progress  over   time, and there will be other people who will include the  picture.

Top of the list

But    VMware was at the top of that list of technologies that we have been    using internally for several years, been very happy with. Based on our    historical relationship with VMware and the offerings that VMware  have   in the traditional VMware space, plus the cloud offerings, things  like Cloud Director and other things, that we felt that those were good cornerstone    technologies to make sure we have the greatest chance of success with    few surprises.

And we needed partners to push the envelope,    because we view ourselves as being innovative and groundbreaking, and we    want to do things that are first in the industry. In order to do  those   with better certainty of outcome, you have to have good  partners, and I   think that’s what we found at VMware.

Gardner: What did you learn? Is there any 20-20 hindsight   or Monday morning  quarterback types of insights that you could offer to   others who are  considering such cloud and/or vertical specialty cloud    implementations?

Rubinow: It goes  back to the comments I   just made in terms of choosing your partners  carefully. You can’t  afford  to have a whole host of partners, dozens  of them, because it  would get  very confusing. There's a lot of hype in  the marketplace in  terms of  what can be done. You need people that  have abilities, can  deliver them,  can service them, and can back them  up.

You can’t afford to have a whole host of partners, dozens of them, because it would get very confusing.

Every    one of us who’s trying to do something a little bit different than  the   mainstream, because we have a specific need that we’re trying to    service, has to go into it with a careful eye towards who we’re working    with.

So I would say to make sure that you ask the right    questions. Make sure you kick the tires quite a bit. Make sure that you    can count on what you’re going to implement and acquire. It’s like    implementing any new technology It’s not unique to cloud.

If    you're leading the charge, you still want to be aggressive but it’s a    risk management issue You have to be careful what you’re doing    internally. You have to be careful who you’re working with. Make sure    that you dot your I’s and cross your T’s. Do it as quickly as you can to    get to market, but just make sure that you keep your wits about you.

• VMware's vSphere 5 Hits the Streets
• VMware Launches a Developer Edition for Cloud Service
• Webinar: Analysts Plumb Desktop as a Service as the Catalyst for Cloud Computing Value for Enterprises and Telcos
• IT Financial Management Provides Required Visibility into Operations to Reduce Total IT Costs
• Want Client Virtualization? Time Then to Get your Back-End Infrastructure Act Together