Wednesday, August 29, 2012

Performance management tools help IT services provider Savvis scale to meet cloud of cloud needs

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

The next edition of the HP Discover Performance podcast series highlights how cloud infrastructure and hosted IT services provider Savvis has been able to automate out complexity and add deep efficiency to its operations.

Using a range of performance, operations orchestration and Business Service Automation (BSA) solutions from HP, Savvis has improved its incident resolution and sped the delivery of new cloud services to its enterprise clients.

To learn more about how they did it, we're joined by Art Sanderson, Senior Manager Enterprise Management Tools at Savvis. The discussion is moderated by Dana Gardner, Principal Analyst at Interarbor Solutions. [Disclosure: HP is a sponsor of BriefingsDirect podcasts.]

Here are some excerpts:
Gardner: What are the main drivers in your infrastructure as a service (IaaS) market?

Sanderson: Savvis is recognized as a global IT leader in providing IT as a service (ITaaS) to many of today’s most recognizable enterprise customers around the world. We offer cloud services and hosting infrastructure services to those customers.

Being an IT department of IT departments, or a dynamic service provider, has a lot of unique challenges that you don’t face in every IT shop that you run into. In fact, we have thousands of customers that we have to support with their own IT departments. So our solutions have to be able to scale beyond what you would find in a typical IT organization.

Gardner: And I should think that efficiency is super-important. It's all margin to you, when you can save and do things efficiently?

Better SLAs

Sanderson: Absolutely. There are just the efficiencies alone for operational cost, as well as the value that we provide to our customers, being able to provide better service-level agreements (SLAs), so their businesses are up and running and available to them to service their own customers. There are definitely some economies of scale there.

Our premier services are our Symphony cloud offerings, our Symphony VPDC, Symphony Open and Dedicated cloud, as well as Symphony Database. All, in some form or fashion in various degrees, use the BSA tools on the back end to do their own offerings, and their own automations that we offer our customers.

Gardner: Tell me what you've done in terms of management for better automation, orchestration, and then, how those benefits get passed on.

Sanderson: Sure. We've adopted the HP BSA set of tools as our automation platform and we’ve used that in a number of different ways and areas within Savvis. It's been quite a journey. We’ve been using the tools for approximately three to four years now.

We started out with some of our operational uses, and they've matured to the point now where a lot of our automation-type monitoring is solved by automation rather than by our operational staff.

There is definitely labor saving there, as well as time savings in mean time to resolution values that we’re adding to our customers. That's just one of the benefits that we’re seeing from the automation tools, not to mention the fact that we build a lot of our own key product offerings for the marketplace that we service, using the BSA offerings on the back end as well.

Gardner: How do you measure performance benefits? Is there a set of key performance indicators (KPIs) or some benchmarks?

In just this first quarter of 2012 alone, we recognized somewhere in the neighborhood of $250,000 in labor savings.

Sanderson: From an operational perspective, we do monitor the number of automations that we run that we can capture from the operational side of the house. For example, on a typical day we run anywhere from 10,000-20,000 types of automations through our systems, and that would actually add value back to the business from a labor-savings perspective.

In just this first quarter of 2012 alone, we recognized somewhere in the neighborhood of $250,000 in labor savings just from the automations from an operational perspective. Again, it's hard to quantify the value of adding to the business side, because those are solutions that we’re offering to the market space that are generating new value back to the organization as a whole.

Mature process

From the people and process side, we didn’t start out necessarily doing it the right way from the operations side of the house. But we have matured the process to where we're now delivering solutions in a much more rapid fashion. The business is driving the priorities from an operational perspective as far as what we’re spending our time on.

Then, we can typically turn around automations in a very short time. In some cases, we’ve built frameworks using these tools where we can turn around an automation that used to take two to three weeks. Now, it can take less than an hour to turn around that same automation.

So we’ve gotten really smart at what we’re doing with the tools, not just building something net new every time, but also making the tools more reusable themselves.

From the value to the organization, we’ve also had many groups within the product engineering side of the house take on and learn tools like HP Operations Orchestration (HPOO) and HP Service Activator (HPSA), and leverage their own domain knowledge as network engineers or storage engineers to build net new solutions that we then turn around and offer to our customers.

That eliminates a lot of the business analyst type of work and things like that that would typically go into the normal systems development lifecycle (SDLC)-type process that you would see. We’re able to cut the time to market for the offerings that we’re producing for our customers.

It does make us much more agile and responsive to the needs of our customers and the industry.

It does make us much more agile and responsive to the needs of our customers and the industry.

Gardner: How large is Savvis?

Sanderson: Today, we have about 25,000 servers under management, spread across 50 data centers worldwide, and just to give you an idea, we have approximately 9,000-10,000 automations on a typical day running through HPOO.

As far as the scale and break down of the servers, two-thirds of our servers today are virtualized, and either through the cloud or actual traditional orders that customers are placing. So, we’re seeing a lot of growth in the virtual machines (VMs) and the cloud space. This is where things are going for our organization as well as the industry.

Self healing

Our self-healing infrastructure is where we’ve actually matured our process and recognized the reusability of using a meta-model to drive our HPOO flows that we’re writing. We've taken those patterns that we’ve identified and have been able to build a meta-model that we now have built a user interface in front of.

If somebody wants a new request, they can go in and request that from us, and then we can, within a matter of minutes, produce the data through the user interface and publish a new flow, without ever having to write new operations orchestrations flows.

Gardner: Tell me a little bit about what your future plans to improve both innovation and productivity?

Sanderson: Obviously, the reason we come to conferences like HP Discover is to learn about where HP is going, so we can make sure that we're in alignment, both from our business needs, as well as where the products are going that we use to drive our own solution.

It's critical that we're able to maintain an upgrade path and we're able to support our business. We've already started to plan, based on what we see coming down the path from HP's future infrastructure and even dedicated infrastructure as our business continues to grow. For example, for the Symphony products that we were referring to earlier, we have to break off more-and-more dedicated infrastructure to the scale and capacity that they’re growing.

We would have never have anticipated, when we started a few years ago, that a customer would have come to us to say that we want to order 400 VMs or we want to order 1,000 VMs, but customers are coming us today doing that. That's the kind of scale that we’re seeing, even just a year into the offerings that we’re providing to the marketplace.
Listen to the podcast. Find it on iTunes/iPod. Read a full transcript or download a copy. Sponsor: HP.

You may also be interested in:

Tuesday, August 28, 2012

Learn why success greets NYSE Euronext's Community Platform for Capital Markets cloud

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

Get the latest announcements about VMware's cloud strategy and solutions by tuning into VMware NOW, the new online destination for breaking news, product announcements, videos, and demos at:

Our next VMworld case study interview revisits a unique vertical industry cloud -- NYSE Euronext's Capital Markets Community Platform -- to take stock of how mission-critical cloud services are being delivered.

We'll learn about how this innovative cloud and groundbreaking business model targets the needs of Wall Street IT leaders, how the business of the financial services industry has received them, and explore how providing cloud services as a business has evolved.

This story comes as part of a special BriefingsDirect podcast series from the 2012 VMworld Conference in San Francisco the week of August 27. The series explores the latest in cloud computing and software-defined datacenter infrastructure developments.

Our guest is Feargal O'Sullivan, the Global Head of Alliances at NYSE Technologies. The discussion is moderated by Dana Gardner, Principal Analyst at Interarbor Solutions. [Disclosure: VMware is a sponsor of BriefingsDirect podcasts.]

Here are some excerpts:
Gardner: How have things progressed over the past year?

O'Sullivan: We've been very happy with the progress we've made. When we announced at VMworld last year, we had just gone into early access for our first clients in our data center in the New York, New Jersey, Connecticut tri-state area, where we have all of our US-based markets running the New York Stock Exchange Markets, the Arca Electronic Markets, and AMEX.

That has since gone into production, has a number of clients on it, is being perceived very well by the community, and is really driving as a lynchpin of our strategy of building a global capital markets community.

Since the success of that, we've actually progressed further, to the point of having deployed the same environment in a second data center that we own and run just outside of London, in a town called Basildon, which is where we run all of our European markets, the Euronext side of NYSE Euronext.

We now have an equivalent VMware-based cloud environment and a range of ancillary services for the capital markets industry available in that location. Clients can now access, as a service, both infrastructure and platform capabilities in both of those facilities.

Furthermore, we've extended to two other financial centers in the world, one in Toronto and one in Tokyo. That's a slightly more stripped-down version of the community platform, but it's very useful for clients who are really expanding the business and gone globally.

Four locations

Now, we have those four locations up and running in production with production clients, so we are very happy with that progress.

Gardner: What is it about the way that we're doing things now -- the whole software-defined datacenter model -- that's allowed you to build out so quickly?

O'Sullivan: Clearly, the technology has advanced significantly from the old days. The capability around virtualization on the the hardware server level with the VMware hypervisors, and in particular the vCloud service suite, gives clients their own control over their environment.

Also on the networking side, it's become much more viable for clients to actually deploy into shared environment, still maintaining confidence that they're going to get both the security profile that they're looking for, as well as the performance capability.

We use the EMC VNX array with the FAST Cache capability to give a very stable performance profile based on demand. It allows different workloads, and yet each gets very good performance and response time. So there are many components along the way. Also, management and monitoring of these types of infrastructures have improved.

Our clients have certainly seen that enhancement in the technology. The financial services industry is unique in the way it leverages technology on two aspects.

One, security profile is absolutely critical. Security isn't just around customer data, but around application development and tools of the trade, intellectual property that firms might have, trading strategies, different analysis, analytics, and other types of components that they develop and build,. They feel they're highly proprietary in nature and don't want to allow anybody to get access to them. So they place security extremely high on the list.

The other unique aspect is performance aspect. It's a slightly different performance model from your typical sort of three-tier web store type of environment. Financial services, first of all, push very high volumes of content through their applications. They need to do so in microseconds, or at least milliseconds, of response time and latency measurements, and they also most importantly need to do so predictably.

With a big batch job of some kind, say a genetic folding job, you drop off a job, go away for 12 hours, and you come back. A little bit of clearly inefficient processing time is not great, because that drags out the whole thing over time, but there is no sort of critical "need it here," "need it now" requirement. So latency spikes are less of a problem.

Latency spikes

But in our industry, latency spikes are a real problem. People look for predictive latency, so we had to make sure that we applied a very tight security profile to our cloud, and a very high performance profile as well.

Gardner: How have you been able to build on this cloud in terms of those value-added services that you deliver specifically to a financial clientele?

O'Sullivan: That's why we built our cloud, because there are many service providers who offer very valuable cloud capabilities that are based on core infrastructure and core computing capabilities, and they do so very well. However, we consider ourselves a vertical industry community. We're specifically focused on capital markets participants. We try to support and make it cheaper, more cost-effective, and more readily accessible to a wider range of participants to be able to get access to the markets.

So in our cloud and our community, we provide a range of platform and services that we have added. The core is "Come into our vCloud Director environment and access your compute infrastructure." By the way, we have a Compute On Demand Virtual Edition, we also have a Compute On Demand Physical Edition for those cases where that latency issue is of the utmost importance.

Then, we provide clients with the value-added features that we know they need, because they're in the capital markets business. The key one is market data. This is something that is absolutely critical in financial services, because every trade, no matter what you are buying or selling, always starts with a quote. Even if you walk into the shop and you ask how much it would it be for a can of soda, they say it's $1 or $1.20, whatever it is, and then you decide if you want to buy.

So in the financial services industry market data is the starting point, the driver of all the business. And the volumes on this, the sheer size of the content that comes down, is really outstanding. It's at the point now that even if you were to just subscribe to all North American equities and options, you'd need a 10-gigabit Ethernet pipe, and at points during the day, you're probably using upwards of 8 gigabits of that pipe just to get all that content.

Obviously, we can provide raw content, but we've added a range of services into our cloud and into the community. We can say, "We can offer you a nice filtered market data feed, where you just present us with the list of instruments you want, and we can add value-added calculations, do analytics, and provide that to you."

We've also developed an historical market-data access service. So if you want to go back and test your strategies against previous days of trading, back for many, many years, we have a database that's deployed in the cloud. So you can query the database, load it into your virtual environment, and analyze and back-test your strategies.

We've added order-routing capabilities, so when you are ready to send your orders to the market, if you are a market maker yourself, you might go direct to our gateway. If you're a sponsored participant, you might go through our risk-managed gateway, which would be sponsored by a broker.

Or if you are just a regular buy-side firm, a money manager, you might use our routing network and ask us to write your orders to the different brokers or the different markets, and we can handle that. Those are either ends of the trade.

Integration pieces

On Thursday, Aug. 30, I'm going to be presenting with VMware and EMC in one of the breakout sessions about us moving up the stack to start offering more of the integration pieces of this. We're using the Spring environment and a range of other VMware tools, GemFire, and so on, to demonstrate a full trading system deployed in the virtual environment with the integration tools -- all running hosted in our environment.

It's more of a framework that we're showing, but it provides platform as a service (PaaS), not just the market data in, which is our specialty, and the order routing out. Once you're within your environment, the range of additional tools makes it easy for you to develop and customize your own trading tools and your own trading strategies. That's something I will be talking about on Thursday.

Gardner: How has the reception in the market?

O'Sullivan: The good news is that we've definitely had great progress here. We have a number of clients in all of the locations I mentioned. We're continuing to grow. It's a tough environment, as you can imagine, both just in the general economy and in particular in the financial services industry. So we expect to continue to grow this significantly further.

We have been certainly very happy with the uptake so far. We knew that we were going out well ahead of everybody else and we were very keen to do so, because we see and understand the vision that VMware and EMC in particular have been promoting over the past few years. We agree with it fully. We feel like we're uniquely positioned within the capital markets industry as the neutral party.

Remember, we're just a place where people go to trade. We don't decide what you buy or what you sell or how much it should be. We just provide the facility, the rules, and the oversight to ensure an orderly market. We wanted to make it easier and more cost-effective for firms to get access to that environment.

So by providing all of this capability, we think we're in a fantastic position now, that as more and more firms continue to explore virtualization and outsourcing of non-business critical functions, which for a while used to be running on your own servers, but which are now nothing but overhead.

We see them moving more and more into the cloud. We expect over the next two or three years, that this is really going to explode. We intend to be there, established, fully in production, tried and tested, and leading the industry from the front, as we think we should be with the a name like the New York Stock Exchange.

Well-known brand

That’s a brand that's so well-known globally. It's the best place to trade. It's the most reliable and most secure place to trade stocks, with the best oversight, and we want to apply that model to all of the services that we offer our clients.
Listen to the podcast. Find it on iTunes/iPod. Read a full transcript or download a copy. Sponsor: VMware.

Get the latest announcements about VMware's cloud strategy and solutions by tuning into VMware NOW, the new online destination for breaking news, product announcements, videos, and demos at:

You may also be interested in:

HP expands Converged Cloud portfolio to work with VMware's new vCloud Suite 5.1

HP today announced at the VMworld 2012 conference in San Francisco that it's expanding its HP Converged Cloud portfolio with new solutions for VMware vCloud Suite 5.1, enabling clients to transform traditional virtualization deployments into private and hybrid cloud environments with less risk and complexity.

Combined with VMware vCloud Suite 5.1, HP CloudSystem doubles VM density, reduces network complexity, and provides added flexibility to support almost any workload, says HP.

Introduced in April, HP Converged Cloud extends the power of the cloud across infrastructure, data and applications. Based on a single architecture, HP Converged Cloud helps integrate many combinations of private, managed and public clouds, as well as traditional IT, providing workload portability as well as dynamic scaling. [Disclosure: HP and VMware are sponsors of BriefingsDirect podcasts.]

The expanded HP Converged Cloud portfolio from VMworld enables clients to:
  • Reduce complexity of building and managing cloud with integrated solutions.

  • Accelerate cloud adoption with comprehensive security and compliance controls that can be automatically applied to virtual machines (VMs).

  • Decrease the risk of data loss with automated policy-based data protection.

vCloud Suite 5.1

The HP announcement dovetails with VMware's unveiling of vCloud Suite 5.1, which delivers its software-defined datacenter, an architecture for implementing cloud computing. vCloud Suite 5.1 extends the benefits of virtualization to every domain in the datacenter – compute, storage, networking, and the associated availability and security services, says VMware.

The software-defined datacenter architecture abstracts all hardware resources and pools them into aggregate capacity, enabling automation to safely and efficiently dole it out as needed for applications. Tenants or customers utilizing the software-defined datacenter can have their own virtual datacenters with a logically isolated collection of all the virtual compute, storage, networking and security resources they are used to.

As organizations move from a virtualized environment to a private or hybrid cloud, they may face significant challenges.

On Tuesday, VMware CTO Steve Herrod unveilved an enterprise mobility solution suite called Horizon. It goes a long way to helping enterprises deliver and control applications and data that can be accessed via most smartphones, tablets and PCs.

New HP CloudSystem capabilities integrated with the VMware vCloud Suite 5.1 include:
  • HP Cloud System optimized for VMware, which provides faster time to revenue with pre-integrated solutions that include HP Converged Infrastructure and cloud management from HP and VMware.
  • Auto-flexing feature, which reduces response time to business requests and improves overall service levels by enabling VMware administrators to provision hardware for vCloud Director to automate the scaling of converged infrastructure.
  • VM importing, which simplifies clients’ path to the cloud by easily importing running VMs into an HP CloudSystem without downtime or reconfiguration.
  • HP Cloud Maps for VMware, which accelerates the design of cloud services and automates the deployment of virtual applications using pre-packaged, optimized HP CloudSystem templates.

Data protection enhancements

Also, at VMworld, Autonomy, an HP Company, announced a major feature enhancement to HP Data Protector 7 that delivers integrated security for vCloud Director 5.1, providing fully automated, policy-based protection for cloud environments.

This is a key offering from Autonomy and provides policy-based data protection for an array of configurations within a virtual datacenter. It also allows organizations to control how often they back up their data, as well as how much data to retain and store.

The solution enables service providers to offer flexible protection options within multi-tenant cloud environments. Organizations that use vSphere and seek to maximize storage efficiency for VM backup can leverage Data Protector 7 to de-duplicate data anywhere – at the application source, backup server, or target appliance – and centrally manage the de-duplicated data movement across the enterprise for disaster recovery.

The solution enables service providers to offer flexible protection options within multi-tenant cloud environments.

Using Data Protector 7, customers can eliminate protection gaps found in cloud-based deployments. Traditional data-protection solutions often fail to understand how much data to back up and the level of data protection required for the diverse types of data that reside within an organization.

Customers can also reduce the risk of data loss, as business critical backup policies are automatically applied to new VMs within a virtual datacenter. And with HP StoreOnce Catalyst replication feature, customers can attain back-up at speeds of up to 100 terabytes per hour, three times faster than the leading industry storage alternative, says HP.

PCI industry security

HP has also partnered with VMware to offer a converged payment card industry (PCI) Compliance Security Solution Stack that provides a single security framework to address diverse virtual and cloud infrastructures. The solution provides comprehensive, cost effective and easy-to-manage protection that enables strong security, beyond simply meeting compliance requirements.

The PCI Compliance Security Solution Stack is a pre-certified reference architecture for testing applications, defending against cyber threats, as well as providing enterprise-wide visibility across virtual and cloud infrastructures. The solution accelerates the path to compliance covering 110 of the 200 PCI requirements, spanning all 12 regulation sections.

The solution leverages market-leading technology from HP TippingPoint, HP ArcSight, HP Atalla and VMware cloud infrastructure, allowing users to:

The solution accelerates the path to compliance covering 110 of the 200 PCI requirements, spanning all 12 regulation sections.

You may also be interested in: