Data Center Fabric Case Studies

January 31st, 2012

This post was written by Jon Toor

Three new data center fabric case studies. Read how IT managers have increased VM density, virtualized more apps, and saved on power, space and cooling with data center fabrics.

Read these for a quick look at the real-world benefits of a data center fabric.

  • xsigo data center fabricRisk management firm runs over 100 VMs per host, boosts performance 50% Read more
  • Law firm increases VM density to avoid a data center expansion Read more
  • Healthcare service provider reduces infrastructure by 77% Read more

While others talk about their promises, Xsigo solves today’s most challenging virtualization problems with data center fabric solutions.

Bookmark and Share

Intel Acquires InfiniBand: Do They See The Fabric of the Future?

January 24th, 2012

This post was written by Jon Toor

Why is Intel getting into the InfiniBand business? This is big news; bigger than the size of the acquisition would suggest.

Certainly it signals Intel’s growing interest in the infrastructure business. But perhaps more interesting is what it says about InfiniBand.

While InifiniBand is best known for HPC applications, I suspect Intel may have its sights on a much broader market than university labs. InfiniBand incorporates a number of technologies that make it very well suited for the enterprise.

Several current data center trends highlight why InfiniBand is well-positioned as the fabric of the future:

1) Server virtualization: To run a large number of VMs, virtualized servers need a lot of bandwidth. InfiniBand delivers by far the most cost-effective, power-efficient performance.

2) Convergence: The only solution for the growing connectivity needs of virtualized servers is to converge the infrastructure. It’s not hard to see the benefit of two cables vs 20 cables. InfiniBand provides the ideal transport for this. It is and always has been a reliable connection. You can run both Fibre Channel and Ethernet traffic over it without packet reordering.

3) Data Center Fabrics: The next challenge for the data center is to fix the East/West problem. Multiple Ethernet vendors are proposing solutions that simplify and accelerate server-to-server communications. Well, InfiniBand was designed as a fabric. The networking challenges faced by Ethernet were addressed in the original InfiniBand design.

Even the knock that InfiniBand is “just for HPC” is not true.

Xsigo has offered for years a data center fabric that converges FC + Ethernet server connections to a single InfiniBand link within the rack. The top-of-rack device fits right in with conventional data center core switching gear. It’s successfully deployed in hundreds of data centers today. And, Oracle uses InfiniBand as the cluster interconnect of their flagship database machine, Exadata.

Intel would probably not undertake a strategic move like this simply to better serve the HPC market. The big prize is the market for servers and infrastructure in next-generation cloud data centers.

Server virtualization dramatically altered the way we look at processors. The sky’s now the limit when it comes to multi-core / multi-socket implementations.

Infrastructure virtualization is the next big wave. Data center fabrics are the products riding that wave. Like virtualized servers, fabrics too will be built on significantly faster underpinnings. With an InfiniBand offering, Intel may be positioning themselves as a leader with a proven technology that was originally designed for exactly that application.

 
 
Bookmark and Share

Adapt or Die: The Evolution of Data Center Architecture

January 17th, 2012

This post was written by Jon Toor

Environmental changes don’t kill off a species. It’s the lack of adaptation that kills. This is the challenge facing IT managers today. The data center fabric can help.

Jon Oltsik of ESG wrote an article about the environmental changes in the enterprise data center. They’re truly tectonic. According to their research:

  • 2/3 of enterprises are consolidating (or have consolidated) data centers
  • Nearly 100% of enterprise data centers now have some level of virtualization
  • Most have aggressive plans for much higher virtualization workloads

Gartner concurs, reporting that the average number of workloads virtualized is over 50% now, going to 75% within two years.

Think back just a few years and you’ll recall a very different picture.

Four years ago, VMware was known largely as a great test and dev tool. Just two years ago, many shops still considered server virtualization suitable only for Tier 2 and Tier 3 applications.

Now virtualization usage is everywhere, but the data center management model remains in some middle ground.

ESG asked managers to identify their management pain points. Some challenges identified included:

  • Too many manual processes
  • Time consuming network provisioning and configuration
  • Organizational problems between the networking team and other functional IT groups
  • Network security

These are exactly the problems that the data center fabric was created to fix.

Your opportunity is to capitalize on these complementary trends — server virtualization and the data center fabric. Together they will provide an extreme makeover of data center economics.

Failure to act could result in… well, we only have to look back a few million years for that analogy.

 
 
Bookmark and Share

Case Study: Service Provider Boosts Data Center Agility, Cuts Costs with Xsigo

January 9th, 2012

This post was written by Jon Toor

Data center agility sounds great, but what does it mean? And how do you achieve it?

This case study from Solutions Healthcare looks at both of these questions. In their business, agility has specific meaning for SLAs and cost. Implementing an agile infrastructure let them increase effectiveness on both of these metrics.

But before diving into that, let’s consider the question of what agility means. It’s a vague, over-worked phrase, and it’s hard to quantify.

But its impact is neither vague nor intangible. In fact, data center agility can make a huge difference in your productivity and your daily life.

If you need to resolve an outage, it could make the difference between minutes or hours of downtime. Or if an update is required, it could make the difference between a  normal work hours task or a middle-of-the-night project.

So how can agility help? Here are six questions to ask.

  1. Can you add servers quickly and with minimal disruption when you need them?
  2. Are you able to easily move VMs from any server to any server, not just within a specific cluster?
  3. Does your infrastructure allow you to run any application on any server?
  4. Is it easy to resolve application performance issues without physically connecting new networks?
  5. Can you routinely resolve outages within the time allowed in your SLA?
  6. Can you launch a new application quickly, regardless of its network and storage requirements?

If the answer to all of them isn’t “yes,” a more agile infrastructure could help.

Jeff Bills, VP of IT at Solutions Healthcare, had exactly these concerns. And he implemented a Xsigo virtualized infrastructure to let him deliver higher service levels to his customers while reducing his own costs.

VMware + Data Center Agility

Jeff got agility in his data center the same way he got it on his servers: virtualization. A virtualized infrastructure decouples the connectivity from the physical connection. You can change the way a server is connected to networks and storage in software, which is a lot quicker than messing with I/O cards, cables, and switch ports.

It’s analogous to server virtualization. Just like you can carve a single processor into multiple independent VMs, infrastructure virtualization lets you carve a single cable into multiple independent links.

Jeff, a long time VMware user, found that putting these two technologies together made everything faster and more efficient. Check out his story for more.

So what do you think? Would more agility a plus for what you do? I’d be interested in hearing.

 

 

 

 

 

 

 

Bookmark and Share

Seven Reasons Why the Data Center Fabric will Take Off in 2012

December 21st, 2011

This post was written by Jon Toor

In 2011, we saw the hype around data center fabric. But things are just getting going.

In a Network World article on converged infrastructure, Analyst Zeus Kerravala said, “If the migration to fabrics was a baseball game, then the pitcher hasn’t even finished warming up. In 2012 however, we’ll see solutions take off.

data center fabric efficiency

Xsigo customer Investec demonstrated data center fabric benefits with operational savings.

Why? First, the need for convergence with fabrics has become increasingly clear. IT managers now run more and more apps on VMs, and at far greater densities.

At Xsigo, we have multiple customers now running 140+ VMs on a single physical host. This puts tremendous demand on the I/O infrastructure, often requiring more than traditional I/O can deliver.

Second, going into 2012, the benefits of convergence are proven. Trailblazing IT shops have paved the way for others to follow.

For example, in the same article Xsigo-customer Investec reports a 90% reduction in network complexity with Xsigo. And they said that server deployments now take just two hours, rather than three weeks, a massive operational improvement. As for application performance, they report it’s better than ever.

Third, we are seeing fundamental cracks in today’s networking model when data centers become heavily virtualized.

How Data Center Fabrics Address Networking Challenges

For a variety of reasons, conventional networks simply don’t meet the needs of virtualization. Here are seven ways that data center fabrics address those challenges.

1.     Relieve the east-west traffic problem

Gartner claims that 80% of data center network traffic now travels from server-to-server. Depending on where it’s going, that data may travel through multiple layers of infrastructure, creating latency and I/O congestion at choke points along the way. This is referred to as the “east-west” problem. Data center fabrics will address this with smarter data paths that can traverse the topology much more directly, adding performance and reducing physical complexity.

2.     Eliminate the spanning tree issue

The Spanning Tree Protocol dates back almost 30 years. Designed to ensure a loop-free topology, it was not intended to build an efficient virtual data center. In fact, it creates major performance challenges by limiting data paths in ways that create congestion. New solutions employ greater network intelligence to route data in faster, more efficient ways.

3.     Reduced reliance on VLANs

VLANs, which also date back to the 1980′s, segregate network traffic. They also saddle virtualization managers with tedious I/O configuration tasks and present scaling challenges in large data centers. Data center fabric will enable simpler, more scalable segregation without reliance on VLANs.

4.     Connect your VMs in seconds

VMs are quick to deploy but not always easy to connect to other resources. You need a network identity and a data path, both of which will require coordination across multiple teams. Data center fabrics will let you connect quickly, especially when connecting to other VMs on the same fabric. In that case, traffic can traverse the fabric independent of the general Ethernet production network, thus reducing congestion and eliminating a host of configuration tasks.

5.     More bandwidth where you really need it: at the server

Traditional networking has skinny pipes down at the server level and progressively fatter pipes as you move up the networking stack. But if 80% of the traffic is server to server, you actually need more bandwidth down low in the infrastructure. In a way, this turns the typical bandwidth pyramid upside down. Putting more bandwidth down at the server level with a data center fabric will help you run more VMs per server.

6.     FC/Ethernet convergence

Most Fibre Channel shops still use dedicated FC server connections to each server, despite the availability of effective solutions that converge Ethernet and Fibre Channel to a single cable. It has been well proven that FC and Ethernet can deliver great performance and reliability when combined over a 20 or 40G link. In 2012 there will be multiple options to achieve full convergence and save serious money.

7.     Open solutions for an efficient path forward

Finally, there will be options to achieve all of these goals while fully leveraging the gear you have. With some of the new data center fabric solutions, this transition absolutely can be smooth and incremental, not rip and replace.

Bookmark and Share

Secure Multi Tenancy: How Xsigo Enables Network Isolation

December 12th, 2011

This post was written by Jon Toor

Secure multi tenancy, always an important concern for service providers, is about to become an even larger issue. Forbes today highlighted draconian new data protection regulations proposed in the EU that will demand even greater care. So how do these new rules bite, and what can you do to avoid getting bit?

In the Forbes article, the author states, “In the first revamping of the EU’s data protection legislation in over fifteen years, businesses would face fines of as much as 5%  per cent of their global turnover (i.e., the ratio of annual sales to inventory) if they are adjudged to have mishandled the personal data of customers, suppliers, or employees.”

secure multi tenancy

Virtual connections are mapped to specific physical ports. Servers can share a common converged fabric and maintain physically isolated links to external networks for ensured secure multi tenancy.

That’s huge. It’s no longer in the realm of just compensating people for damages. These are penalties that could apply to any violation. And they can apply even in the face of determined hackers who cleverly find a way in.

Secure Multi Tenancy

One way to protect information is to keep it isolated within a walled-off environment. In other words, put that information only on a network that never gets out of the building. Creating such physically isolated networks is a well-established practice.

The advent of virtualization, however, adds a new challenge for secure multi tenancy: we want fixed, isolated networks for data protection, but we want data center agility to get the most from virtualization. How do you reconcile those two, sometimes-competing needs?

Xsigo virtual I/O provides a very simple answer. It lets you deliver isolated network and storage connections to servers, just as you would with physically separate networks. But you can add and delete those connections to any server at any time, entirely in software.

Xsigo’s technology is not the same as using VLANs that share a common Ethernet network. Xsigo’s virtual connections have been verified in an independent security audit to be fully isolated. In fact, companies already use it to isolate medical records, payment card information, and financial data.

For a quick look, here is a secure multi tenancy data isolation solution note. If you would like a copy of the audit, please drop me a note at jtoor “at” xsigo.com. Or learn more about the Xsigo I/O Director here.

Physically isolated networks with secure multi-tenancy within a virtual environment. Now you can have your cake and keep people from stealing it, too.

 

 

 

 

 

Bookmark and Share

Xsigo Showcases Power Savings in VMware’s Tokyo EBC

December 6th, 2011

This post was written by Jon Toor

VMware’s Executive Briefing Center in Tokyo highlights the latest in virtualization technology, including the Xsigo I/O Director.

Demand for virtualization in Japan surged when recent power shortages spurred data centers to find new ways to save power.

Xsigo virtual I/O can save about 100 watts per server by consolidating I/O cards and reducing switch ports by 70%. It also permits the use of smaller, more power-efficient servers by eliminating the need for multiple PCI slots.

Here’s a photo of VMware’s demo rack.

 

Bookmark and Share

Data Center Agility (or, Two Weeks Work Done in Five Minutes)

December 5th, 2011

This post was written by Jon Toor

“Data center agility” sounds great, but what does it really mean? We recently saw a great agility example with a Xsigo customer, Farrer & Co. This law firm saved about 99.9% of the time required to complete an important reconfiguration project, finishing the job in minutes rather than weeks.

When Farrer & Co., a London-based law firm, embarked on a data center infrastructure upgrade, they targeted increased agility as one of the project objectives. They had an ongoing need for agility because the practice of law is becoming increasingly information technology-intensive. Client requests now have a much higher IT component than in the past. As the firm’s infrastructure supervisor, Peter East, noted, “We get a lot of short notice requirements.”

To help meet their client requests, they deployed Xsigo virtual I/O. Xsigo’s ability to reconfigure I/O on the fly, entirely in software rather than hardware, and on live servers, gave them the infrastructure agility they wanted. Assets can be repurposed in a hurry.

What they did not necessarily foresee was exactly how much time it could save in an important project.

Significant time savings on a VMware project

When Farrer had a pressing need to run their practice management software in a fault tolerant VMware environment, the use of VMware’s FT feature was an obvious solution. The only problem was this: FT requires a dedicated network.

With traditional infrastructure, configuring a dedicated FT network would have taken several weeks. But with Xsigo they had the job done in five minutes.

Farrer deployed a fully isolated Ethernet network, with performance guaranteed by hardware enforced QoS, in less time than it takes to get a cup of coffee.

That’s the meaning of “data center agility.”

This is just one example, but Farrer found that agility helps them every day in providing better service to their clients.

Xsigo support saves time as well

They found that Xsigo’s team was pretty agile too, leading the firm’s IT Operations Manager, Andy Beech, to comment, “In addition to providing Farrer with an agile network environment, the Xsigo team also demonstrated their responsiveness when working with Farrer’s IT department during implementation… The Xsigo engineers understand our issues and respond promptly. It’s refreshing to deal with a very professional team that is always there and always pulls through.”

To read the entire case study note, click here.

Bookmark and Share

Xsigo XMS 3 Demo: Basic Virtual I/O Management

December 1st, 2011

This post was written by Jon Toor

See the basics of Xsigo XMS 3:  Create virtual NICs and HBAs, monitor performance, and deploy templates to replicate I/O configurations.

Try XMS 3 for yourself right now! For access to an online data center simulator, register here.

 

 

Bookmark and Share

Increase VMware Performance 4X with Xsigo

November 28th, 2011

This post was written by Jon Toor

How do you increase VMware performance? ViaWest has innovated in cloud services with an interesting new hybrid offering called their KINECTed Cloud that capitalizes on the performance and any-to-any connectivity of Xsigo virtual I/O.

Jason Carolan, Vice President of Product Development at ViaWest, describes it like this:

increase vmware performance with virtual I/O“ViaWest’s KINECTed Cloud service offers our customers a complete hybrid solution that provides cost-effective managed services for day-to-day needs and capacity-on-demand cloud services to satisfy peak requirements, all within a single services and support framework.”

Read about it in this Network World article.

Xsigo played a central role here by providing the virtual I/O glue. Any-to-any connectivity that allows ViaWest resources and customer resources to be joined as needed.

“Xsigo’s virtual I/O provides the dynamic connection point that allows us to combine resources – our data center capacity plus our customer’s capacity – to address within minutes the customer’s immediate business challenges.  Xsigo offers the flexibility to seamlessly interoperate with all of our customers varied resources and furthermore delivers performance that is 4X faster than10G Ethernet.”

Increase VMware Performance with 40Gb Connectivity

Specifically, Xsigo lets ViaWest:

  • Connect any server to any data center resource in real time
  • Consolidate Ethernet and FC traffic to a single cable
  • Deliver up to 40Gb throughput through each physical connection
  • Accelerate processes such as vMotion for faster, easier management

Or as Jason summarized it, “Xsigo gives us a way to connect into that (SAN FC) legacy, provides a great bridge… (and) lets us have a managed platform to slice and dice with greater granularity than other products out there.”

Take a look at the Network World article for a view of where cloud solutions are headed. For more information on ViaWest and their new offering, visit their web site here.

 

 

Bookmark and Share