Virtualization

Arista continues to drain Cisco’s brain

Arista Networks, a networking start-up that sells high-end Ethernet switches for cloud computing environments, continues to hire former Cisco Systems executives.  At the top is CEO Jayshree Ullal, who joined the company last year after leaving her role as a senior vice president in charge of Cisco’s $10 billion data center, switching and services business.

On her blog today, Ullal revealed two new hires, both former Cisco executives.

Doug Gourlay has joined the company as vice president of market, just a couple months after leaving his role as Cisco’s vice president of data center marketing.

Anshul Sadana has joined Arista as the company’s vice president of customer and systems engineering. Sadana has been with Arista since 2007, but Ullal is just announcing his appointment as an engineering VP. Before he came to Arista, Sadana headed up Cisco’s development team for the Catalyst 4500 and 4900 product lines and also managed strategic customer relations.

Network considerations for virtualizing your SMB

At Burton Group’s Catalyst conference 2009, Matt Lavallee of MLS Property Information Network presented a case study in enterprise virtualization on an SMB budget. Since server and storage virtualization are network problems, I thought I would speak to Lavallee himself to get his perspective on what networking professionals need to do to overcome the challenges of transitioning to a virtualized environment in a small and medium business (SMB).

After his session, LaVallee sat down with me in a video interview to speak to our SearchNetworking.com audience about virtualizing SMBs. In our Q&A, you’ll learn some of the challenges in transitioning to a virtualized environment for SMBs, then understand how network managers and engineers need to change how they view their network to accommodate a virtualized infrastructure. Are there any tools that can help? Find out in this six-minute video:

Computer networking trends 2009 from senior Burton Group analyst at Catalyst conference

The needs of the network are changing in 2009 as enterprises centralize their data centers, rely on fluid storage sources and depend more and more on wireless LANs instead of wired Ethernet. At this year’s Catalyst show in San Diego, I spoke with senior Burton Group analyst David Passmore, to grasp what trends will directly impact enterprise technology and what IT professionals need to prepare for in both the short and long term. Here is what Passmore had to say:

1. As the networking track chair of Burton Group’s Catalyst conference 2009, what major themes will you be addressing this year?

David Passmore: From a networking standpoint, there were four areas we thought would be of most interest for our enterprise IT clients:

• Wireless is one because there’s an increased use of mobile phones for both data as well as for voice. We’re also seeing enterprises using wireless LANs (WLANs) often as a substitute for wired Ethernet.
• The second focus is on wide area networking and what’s happening in the telecom industry, with particular emphasis on what enterprises can do to save money. If you’re a large enterprise, your phone bill, for example, is the second largest recurring expense after salaries. Obviously, if there’s any way for large enterprises to save on their telecom costs, then that’s going to be of interest.
• A third area that we wanted to focus on is unified communications, which essentially is an outgrowth of phone systems — where most large enterprises over the last couple of years have been migrating over to VoIP and IP telephony. Now what they’re doing is trying to figure out how to combine these systems with instant messaging and presence and email, and other forms of communication.
• The fourth area is data center networking: most enterprises have in recent years, consolidated their datacenters: So they’ve gone from a large number of smaller data centers to a small number of larger datacenters. In doing so, now they’re beginning to implement server virtualization. They have these large storage arrays. And essentially, they need a new network that can support the IT requirements in a consolidated datacenter. So we’ll be talking about — for example, how to have your network work well with virtualized servers, and with storage area networks based on Ethernet, and trying to reduce cabling costs — a lot of the issues that affect networks in large enterprise datacenters.

2. How have these themes changed since Catalyst two or three years ago?

Passmore: Each of them has their own specific changes. A big emphasis this year given the economic environment is saving money or cost avoidance. So a lot of the discussion is around initiatives that provide a near-term economic payback.

3. Are the trends this year something we can apply to today, or something to think about for the future?

Passmore: Actually, it’s a combination of both. One of the things we always try to do at Burton is to avoid just telling people what’s worked well in the past. We tell large enterprises what they need to pay attention to or what needs to be on their radar screens as they go forward — such as wireless. The fact that, increasingly, communication is migrating from wired infrastructures to wireless infrastructures, and that includes, not just the use of WLANs within the enterprise, but more use of cellular, data and voice networking.

At some point this might actually make the [networking professional's] jobs easier, because they might find that more and more of their enterprise’s communications is actually flowing over a wireless network operator['s] network. There will be less of a need for enterprises to have to run their own circuits to manage their own networks for connectivity between or within their sites.

Right now, it’s more of a burden because enterprises have to worry about both wired and wireless communications, and people will look at that as doubling their expenses.

In the near term it, it may be doubling the expense, especially because enterprises are interested in how they can get those to work together. For example, [users] have a single phone number that works for both your desk phone and your cell phone; [they] have a single voice-mail box that works for both their desk phone and their cellular phone.

So that’s the near-term goal, but longer term, we expect that a lot of the enterprise equipment will go away in favor of people making use of mobile devices and using the cellular network operator services.

To learn more about these trends, Passmore speaks particularly about how to overcome the challenges storage and server virtualization puts on networking pros, in this Q&A.

Are you a cable boy or a network engineer?

Recently SearchNetworking.com ran a news story highlighting the emerging skills necessary for network engineers and administrators to survive. The article summoned an outraged call (I believe the word he used for his temper was simmering) from a network engineer at a global firm. He was appalled that any network engineer or administrator would not already come equipped with the skills we outlined.

The story in question called on networking professionals to train themselves in application management, cloud computing, security – a lot more than providing simple IP pipe access.

“Opening dumb pipes is for cable boys. If you don’t know this stuff already, you’re not an engineer,” my annoyed friend said. “You’ve set the bar too low.”

My first thought? Do you really believe that all high level network professionals are ready to push into emerging technologies without some urging?

My second thought: What does this guy have against cable boys? And isn’t that as un-PC as you can get in the world of networking?

But that’s not what he meant.

“The assumption of the story should have been that the network is seeing a demand for more senior level network administrators,” he said.

Sound a bit arrogant? Maybe, but he explained that a real network engineer, administrator or manager is a homegrown product that starts as a cable boy and listens to others to learn. Real engineers, he said, don’t get one or even two certifications and settle for what vendors feed them.

As they grow, they find the humility to reach across IT silos and get over networking protectionism to learn more about complex technologies that increasingly sprawl across departments.

At every level of the networking team, a professional will find this is the time to form alliances between systems, security and networking groups.

“If you have one guy who does firewall and that’s all he does, you’re going to have a problem,” he said.

What’s more, if these departments don’t approach the C-level executives together, they are not likely to get the resources and support they need to implement complex technologies.

“If you’re not working as an alliance, you’re setting yourself up to be outsourced,” he warned.

Cisco puts the cloud computing on edge router blade

This week Cisco announced a bunch of new cloud-based security and collaboration products, including a new cloud-based intrusion protection system (IPS) and Cisco ASA 5500 Series 8.2, with a new botnet traffic filter.  The company also announced that is re-branding the WebEx MediaTone Network, a series of eight data centers around the world that make up the WebEx cloud, as the Cisco WebEx Collaboration Cloud.  This cloud network now offers enterprise IT departments policy control over WebEx meetings, empowering IT to set policies about desktop sharing and file transfers. It also offers global load balancing and intelligent routing, making sure that users enter the WebEx cloud through the best ISP to the most convenient and least taxed data center in the network.

However, what caught my eye in this series of announcements was a new WebEx blade device designed for the ASR 1000 router series. The WebEx Node for ASR 1000 basically transforms Cisco’s ASR 1000 router into a node on Cisco’s WebEx cloud.

Here’s how it works: Let’s say a company wants to hold a WebEx training session for 500 employees. In the old days, each of these 500 employees would log onto WebEx individually across the wide-area network (WAN). With the WebEx Node blade, the ASR 1000 router acts as a broker between the users and the WebEx cloud. The blade establishes a single session with the WebEx cloud. The 500 employees connect through the corporate firewall to the ASR 1000 router and the router connects to the WebEx cloud. By having just one connection to the cloud, shared with hundreds of employees, an enterprise can reduce the amount of bandwidth consumed. This will be especially handy when a company wants to stream high-definition video or send voice-over-IP and/orlarge data sets through WebEx. Companies will avoid WAN bottlenecks and employees will enjoy a better user experience.

Cisco and HP: Data center frenemies now poised for all-out war

What would the data center vendor market look like today if Carly Fiorina hadn’t been ousted as CEO of HP back in 2005? Under her leadership, HP maintained a cozy detente with Cisco. The two mega-companies happily engaged data center customers together.  Cisco sold them switches and routers. HP sold them servers, storage and management software. And HP’s networking division, ProCurve, was consigned to operate in a relative backwater, carving out a solid niche with a pipeline into the SMB networking market.

Current HP CEO Mark Hurd has changed things up. First he named Marius Haas, a rising star at HP who had spent the previous four years overseeing the absorption of HP acquisitions, as ProCurve’s new senior vice president and general manager. Then Hurd and Haas snapped up the well-regarded wireless LAN vendor Colubris, giving ProCurve instant WLAN cred. Before the Colubris deal, ProCurve’s WLAN strategy was built upon an OEM partnership with Motorola.

Word soon came down from above. The incentives that HP had long offered to sales representatives who sold Cisco gear along with HP servers and storage were off the table. ProCurve products were the new priority.

Then last month ProCurve announced its first purpose-built data center switches. ProCurve executives made it clear while briefing reporters and analysts about these new switches that the incentives HP sales reps had for selling Cisco products were long gone. HP would be bringing the full might of its data center presence to bear on its ProCurve strategy. Enterprises could now expect HP sales engineers to offer packages of HP servers, storage, switches, software and services. Quite a proposition.

Of course, none of this has been happening in a vacuum. Cisco hasn’t been sitting still. For a couple years now, Cisco has made it clear that it intends to conquer all things data center as well. It has invested more than $1 billion in rolling out its new Nexus switch line. It has unleashed a barrage of new data center management software and services, labeled Data Center 3.0. And rumors continue to buzz about “California,” Cisco’s much anticipated entry into the blade server market.

So what happens next? It’s safe to say this battle will result in some acquisitions as each company tries to add some weapons to its arsenal.  Allan Leinwand at GigaOM recently suggested a whole bunch of acquisition targets for HP.  For instance, he suggested that HP snap up Arista Networks, Blade Network Technologies, or Force10 Networks in order to beef up its 10/100 gigabit Ethernet portfolio. For storage optimization, he suggested someone like DataDomain.  He said HP should expand into WAN optimization and application delivery, by picking up someone like Blue Coat Systems or Zeus Technology. He also suggested HP target one of the emerging cloud computing specialists.

Meanwhile, Ashlee Vance at the New York Times blogged that Cisco is hoarding cash, leading many to speculate that a flurry of acquisitions is on the horizon.  Vance says that Cisco CEO John Chambers is looking to strike next in the consumer electronics market with the $30 billion in cash it has on hand right now. But enterprise vendors are also rumored targets. Given Cisco’s strong investment in expanding its data center footprint, I think it will spend some of that money on vendors who will help it make war on HP. The EMC rumors just won’t go away, for instance. Last year I heard some whispers that Cisco might make a smaller deal for network management software vendor SolarWinds, but I haven’t heard much about such a deal lately. I’ve seen speculation that Cisco might also target VMware, which it already owns a small stake in. That would be a huge deal, but why would EMC sell it? VMware is a big performer for it. Cisco might buy EMC just so it can have VMware, but the price would be steep.

Once the dust settles over the acquisition blitz, what happens next? I just read a great blog post by Christopher Hoff (hat tip to IDC’s Abner Germanow) which offers a great overview on where all of this is going. For instance, Cisco isn’t really getting into the server business, he says.  Instead, the so-called blade server Cisco is rumored to be working on is a natural outgrowth of the convergence of computing, where storage, servers and switches are becoming more tightly integrated into one infrastructure that supports virtualization and cloud computing.  He writes:

My point is that what Cisco is building is the natural by-product of converged technologies with an approach that deserves attention.  It *is* unified computing.  It’s a solution that includes integrated capabilities that otherwise customers would be responsible for piecing together themselves…and that’s one of the biggest problems we have with disruptive innovation today: integration.

I imagine HP plans to travel down this road as well. Indeed, this should be a very interesting year.

How network engineers can avoid the data center move blues

Just because we’re surrounded by bad news doesn’t mean we can’t turn our lemons into lemonade. That’s at least what network management software solution company Advanced Systems Group (ASG) has done (not to be confused with the ASG band). They found a way to help enterprise IT/network administrator’s avoid datacenter move blunders in a way more compelling than a white paper — they sang about it in their YouTube video: The Data Center Move Blues.

Because we all know, when a data center moves, the network can’t go down. “Oh you know it can’t go down,” wails ASG founder and band front man John Murphy on what appears to be a Martin Backpacker Guitar.

When relocation, deduplication and virtualization complicate the network enough to make IT engineers sing the blues, it’s good to know what not to do during the data center move process.

ASG suggests avoiding these bad boys:

• Bad move #1: The so-called professional help
• Bad move #2: The every-man-for-himself move
• Bad move #3: The do-it-yourself move

More importantly, don’t forget to plan a great deal with your data center team and reach across to the right people. Before making a move, you can’t afford not to plan for every risk, and disaster preparedness will help you stay up while everything else is down.

Five things network administrators do to mess up their enterprises’ virtualization

Network monitoring authority Paessler PRTG has seen the good, bad and ugly of network designs that prevent monitoring a system, which is why their product line solutions cover everything from network management to server performance.

To get an idea of what they’ve seen, PR rep Michael Krems got the company talking about the top five most common mistakes network administrators and IT systems managers make that cause their virtualization deployment to fail.

Here’s what Paessler says will mess up your enterprise’s virtualization:

1. Virtualizing systems without knowing their usual CPU/memory load, disk usage and network usage: You must monitor a system prior to virtualization in order to know how much load it will put on your VM host servers. System with high load may also be not suitable for virtualization at all.
2. Running too many VMs on a host: causing overloads: All virtualized systems suffer in performance.
3. Running too few VMs on a host: spending too much money buying too many host server[s]
4. Compare mid/long-term monitoring results before and after virtualization to ensure quality of service doesn’t suffer.
5. The performance of all virtual systems on a host usually suffers from one virtual system going amok or running into a performance/load peak. Without monitoring, such events often happen undiscovered.

It seems like applying common sense would fix a lot of the issues surrounding virtualization. I wouldn’t place a lot of blame on the network manager though. Much of what goes wrong with a virtual deployment just has to do with the capabilities of the technology. Take these issues for instance:

• Server virtualization impacts network latency.
• Server virtualization creates a network configuration burden.
• Virtual machines present dynamic environment issues for network pros.

As with any new technology, most of the struggle too, is not having the right information. In the end, do you feel like enough conversation surrounds the impact virtualization has on your network?

Riverbed spills into storage market

Riverbed, a well-known vendor in the WAN optimization space, is now expanding their product line to include a new storage optimization appliance called Atlas — which will place them as a strong contender in the storage market space. Although the product is not set to launch until the beginning of next year, it is said to revolutionize data storage by applying some of its WAN optimization tricks to the server.

How will Atlas put them on the map? The concept of the technology is as VP of Product Marketing Alan Saldich says: “It’s almost like virtualizing data.”

For those unfamiliar with their typical WAN optimization deployment, steelhead appliances are applied to either side of a WAN link to conduct deduplication in storage file servers. But Riverbed’s Atlas steelhead appliances will break files down into bits, and send only information to the server that hasn’t been seen before.

“The Atlas appliance is very much like [an] indexer. It doesn’t hold the information; it holds pointers to where that information is stored,” says Saldich.

Since so much of back-up data is duplicated, Riverbed finds their method would potentially save as much as 80% of storage space in file servers.

What does this mean for their WAN optimization products? Although some may see this product announcement as a way for Riverbed to drift away from their original market, they are merely diversifying their portfolio. PR rep Nicole Schlossberg says “Riverbed is not leaving the WAN optimization space. Atlas is an addition to the product line.”

The line between network and storage markets is too permeable for a vendor to see them as two separate spaces. It makes sense that a WAN optimization vendor could also optimize your storage. And though the marketplace has dissolved the line between these two siloed groups, the workforce hasn’t moved its IT structure to adapt.