The stats are in: More bandwidth does not fix application performance problems, according to KillerApps 2013 — a recent Vanson Bourne survey of 650 US and European IT decision-makers.
Respondents were asked if they added bandwidth in the past year, as well as whether they encountered an increase in application performance issues. As it turns out, 92% of U.S. companies are adding bandwidth by at least 10% per annum. In tandem, 69% of U.S. organizations are experiencing a frequency of application latency or non-responsiveness. Of those experiencing more application slowness, business critical apps (especially CRM, ERP, UC and collaboration systems) are impacted more frequently than ever.
So what’s missing? If bandwidth is being added, shouldn’t application performance issues decrease? The answer is “only somewhat.” If organizations add more bandwidth without any application control, it won’t stop one application from usurping bandwidth from a more-important application traversing the network. Solving application problems is not just about adding more bandwidth; it’s about adding network traffic management and traffic prioritization via quality of service (QoS), said TRAC Research President Bojan Simic.
“Monitoring user experience for VoIP and collaborative apps is a good way to measure application performance,” Simic added. Not only is it important to control your applications, he said it’s important to have the proper metrics for measuring such application performance.
Let’s also not forget that WAN optimization controllers (WOCs) and application delivery controllers (ADCs) are proven winners at accelerating app performance. So before you add more bandwidth to your network — consider better network metrics, a QoS strategy and a WOC or ADC to better the application performance experience for your users.
A year ago, IT performance vendor Riverbed Technology claimed competitor Silver Peak Systems had infringed upon five patents.
Just yesterday, the United States Patent and Trademark Office (USPTO) granted Silver Peak’s requests to reexamine two of them (7,428,573 & 7,849,134) and issued Office Actions rejecting all claims in each patent Riverbed accused Silver Peak of infringing — according to an official press release.
The two patents originate around Riverbed’s core deduplication technology. It’s not the first time Riverbed has been in the courthouse over patent infringement. The case is reminiscent of Riverbed’s deduplication debacle in 2007-2008 where backup vendor Quantum sued Riverbed for patent infringement.
Silver Peak CEO Rick Tinsley says Riverbed’s scaretactics call into question Riverbed’s own patents:
It appears Riverbed’s attempts to use its patents to slow down Silver Peak has backfired in a big way. This is a huge embarrassment for Riverbed as it calls into question the originality of their foundational technology. Despite spending millions of dollars on legal fees, Riverbed has to-date been unable to block Silver Peak who is gaining market share with its superior technology.
Although Riverbed is continuing growth, Silver Peak isn’t the only competitor encroaching on marketshare from Riverbed and other dominant WAN optimization vendors (like Cisco). Exinda and other innovative WAN optimization vendors are finding ways to more-flexibly deliver traditional WAN optimization technology.
While Riverbed coined the term “WAN optimziation” and proliferated the technology with its Steelhead devices — the traditional hardware-at-both-ends method of delivery is no longer dazzling customers. Enterprises want software or cloud services to take care of optimizing applications.
This is not to say that Riverbed hasn’t gone through great lengths to keep up with the times. After all, the company has been gaining newer capabilities through acquisition, like cloud networking with its Zeus Technology acquisition or Web content optimization with Aptimize.
The main difference between the competing companies is that Riverbed started out by accelerating key business applications, while Silver Peak began by optimizing anything over IP. Now that applications are proliferating networks, Riverbed has repositioned itself as an IT performance vendor (rather than a WAN optimization vendor) in order to “deliver performance of all types of workloads,” Riverbed Products and Marketing President Eric Wolford explained.
Time will tell which company’s strategy will win — in and out of the courthouse.
The best WAN advice of 2012
What were the most popular tips for network engineers and IT managers dealing with enterprise wide area networks (WANs) in 2012? We took the top five advice articles from SearchEnterpriseWAN.com to give you a countdown of the year’s most pertinent tips:
It’s no secret mobile devices have infiltrated enterprise networks, whether through the consumerization of IT or through BYOD policies. In turn, IT managers have seen more network traffic coming from mobile applications. How can IT calculate the bandwidth requirements for these mobile applications? Tip writer Michael Finneran takes you through the calculations in this top tip of 2012.
Despite the hype around video conferencing and other real-time based applications, a great deal of network traffic is still based on TCP/IP. In this tip, WAN optimization expert Henry Svendblad offers three ways for enterprises to optimize TCP-based network traffic.
As cheap as bandwidth has become, it’s still a huge enterprise expense. Often times, the bandwidth enterprises pay for is not what they are getting in terms of network throughput. That’s why we asked network testing guru Kevin Tolly to give us tips on how to measure wide area network throughput. In this article he shows us how to use free networking tool, iPerf, to accomplish this.
Since the term “virtualization” was coined, the tech industry has gone to great lengths to virtualize just about everything imaginable — including WAN optimization controllers (WOCs). While at first virtual WOCs performed less efficiently than their hardware counterparts, that no longer speaks true today. Nemertes Research Principal Analyst John Burke presents these tips to help IT decide between hardware and software (virtual) WAN optimizers.
More than measuring network throughput, calculating bandwidth and choosing WAN optimization controllers — IT managers have wanted nothing more than to make the right telecom service decision. The most popular WAN connectivity options have included Ethernet and MPLS. Experts Johna Till Johnson and Paul Kirvan explain the benefits and drawbacks of each and list them neatly in this tech comparison chart.
All network protocols were not created equal; nor were they designed for every network type in mind. When it comes to the wide area network (WAN), any issue with a protocol, no matter how small, is magnified as it traverses the great lengths of a wide area.
Take Network File System (NFS) or Common Internet File System (CIFS), for example. These file protocols are extremely latency-sensitive and function poorly over a WAN. TCP/IP (Transmission Control Protocol/Internet Protocol) is certainly not exempt from these inefficiencies.
WAN optimization expert Henry Svendblad explains it this way:
TCP has blunt flow-management algorithms: If the network or the receiver can’t handle the speed at which the sender is working, as indicated by packet loss, timeouts or an excess of out-of-order packets, the network flow will drop to half that rate. Speeds ramp up again, but more slowly than the first time.
Fortunately, IT professionals can take these steps outlined by Svendblad to thrwart slow speeds over a WAN.
For TCP optimization:
- Get smart about congestion and use a WAN optimizer that can change ramping behaviors to prevent congestion in the first place.
- Think about network priority and shape network traffic to control how bandwidth is consumed.
- Keep TCP out of the way of UDP traffic, like VoIP and video, to ensure priority of these real-time applications.
- Consider upgrading to current versions of your enterprise applications that use these more efficient Web protocols.
- Consider using WAN optimizers to shape traffic and control how limited bandwidth is used.
- Use a WAN optimizer to speed up file transfers with compression, caching and prefetching techniques.
For years, the WAN optimization market targeted enterprise customers. In most cases, adding WAN optimization controllers (WOCs) at branch offices cost less than adding more bandwidth. Even still, each WOC would cost anywhere from $2,500 to $15,000 — making it difficult for smaller organizations to justify the cost at every office location.
Some vendors, like Ipanema, have made strides to drive down the cost of WOCs to under $1,000 to target small and medium-sized businesses, but more organizations are looking at other WAN optimization form factors: namely, mobile and cloud optimization.
Mobile WAN optimization clients are ideal for the road warrior or the micro branch office — consisting of up to 10 employees. While a hardware WOC wouldn’t make sense, a virtual WOC doesn’t have a host server to load the software. Most laptops, fortunately, have space for a mobile client, and licenses have dropped from $70 to $30 USD, ZK Research Principal Analyst Zeus Kerravala said.
For those who don’t want hardware, virtual or even software WAN optimization, there is cloud WAN optimization. WAN optimization as a Service has gained popularity in recent years because any number of end users at virtually any location can subscribe to a provider with an optimized link — all with more flexibility and at a more favorable cost.
But neither option comes without its caveats. Nemertes Research WAN optimization expert John Burke highlights the pros and cons of mobile and cloud optimization in this article and even considers how it compares to virtual and hardware WOCs.
Advances in today’s virtual technologies have made deciding between the two form factors more difficult. When virtual WOCs (vWOCs) first entered the market, the technology accelerated data more slowly than their physical counterparts—but that’s beginning to change.
The decision between hardware and software WAN optimizers has become less of a question about speed, and more of a tradeoff between scalability and throughput. It also becomes a question of how familiar IT professionals are with virtual technology. If someone isn’t familiar with a virtual machine or server, then a software WOC could be more trouble than it’s worth.
Nemertes principal analyst John Burke breaks down the hardware vs. software debate this way:
IT needs to balance throughput requirements against scalability, infrastructure availability, and your desire to neatly modularize network functions. Where IT would have to deploy virtual machine hosts in places it did not otherwise plan to, a virtual WAN optimization appliance is not recommended. Where capacity is available and IT wants to spread WAN optimization widely, virtualization has more to recommend it.
To learn more about the advantages and disadvantages of hardware and software WAN optimizers, read Burke’s article on SearchEnterpriseWAN.com.
Life was once simpler for the network architect: Corporate data was located in a data center that belonged to the company, and end users used to work at the headquarters or branch office. If end users complained about a stalling application, a deathly-slow file download or an eternal backup, IT could look into the appropriate WAN optimization or application acceleration solution to make that data run faster. They could put a WAN optimization controller (WOC) at the headquarters and branch and accelerate data between the two controllers — likewise with an application delivery controller (ADC).
But times have changed: Software as a Service (SaaS) makes more sense for organizations; data and storage live in the cloud instead of on-premises; end users don’t just access corporate resources at an office, but also at home and on the go. These changes break down the traditional hub-and-spoke network architecture that once supported older business models. Today, making that data run faster for the cloud and mobile worker is nearly impossible with a hardware WOC or ADC.
Fortunately, vendors have caught on and developed virtual or cloud optimization solutions to keep up with this evolution. Yet cloud WAN optimization is not factored into WAN optimization market research reports, which paint a picture of a fading market. While WAN optimization may be called a different name — like “cloud optimization” or “IT performance solution” — the technology will be there to accelerate any data traversing the network. So when looking for a solution that can accelerate your data — wherever it may live — be on the look-out for non-traditional technology.
As enterprises grow and spread out into branch offices, the ability to move data around becomes critical. Employees need access to information regardless of location, and enterprises need the peace of mind to know their data can be retrieved quickly for remote users, as well as for disaster recovery scenarios.
WAN optimization vendor Silver Peak has been on the cutting edge of data center virtualization for the enterprise, and the United States Patent and Trademark Office recently recognized the vendor’s dedication to distance data movement by awarding Silver Peak with four patents for its Virtual Acceleration Open Architecture (VXOA) software.
The patents include:
- US Patent #8,171,238: “Identification of Data Stored in Memory”
- US Patent #8,095,774: “Pre-Fetching Data into a Memory”
- US Patent #7,948,921: “Automatic Network Optimization”
- US Patent #7,945,736: “Dynamic Load Management of Network Memory”
The software addresses the evolving enterprise WAN requirements — including shrinking real estate in the data center, noted David Hughes, founder and CTO of Silver Peak in a recent release.
In addition to the physical data center changes, enterprises must contend with new struggles associated with moving data across shared or public networks, and especially across long distances.
The newly-patented VXOA software can be deployed on any enterprise server, and move data up to 90 times faster than existing architectures, according to Silver Peak.
The shift from hardware to software isn’t just a data center trend, and adaption will be important for WAN optimization vendors moving forward. As customers are looking toward software-based WAN solutions for their changing business demands, vendors must add virtual WAN optimization software in their portfolios.
If the influx of mobile devices has proven anything, it’s that application delivery and Web content optimization go hand-in-hand. Tablets and smartphones are on the rise, and users need mobility. Traditional application delivery vendors are bracing for the future by aligning their products with the changing needs of the enterprise — faster, and more flexible Web experiences.
Riverbed has announced updates to its Stingray application delivery controller (ADC) product line adding its newly-acquired Aptimizer Web content optimization product to the ADC — Stingray version 9.0.
The announcement comes in at the heels of F5’s latest updates to its BIG-IP Application Delivery Controller (ADC) product line — which was enhanced to support Google’s SPDY protocol and designed with remote and mobile users in mind.
High latency for Web-based applications and slow-to-load websites are the enemy to any business, and the release of Stingray 9.0 sought to extend IT performance in the user direction, noted Naveen Prabhu, senior product manager for Riverbed.
“After just seven seconds, users will leave the Web site if it hasn’t loaded, and these users leaving are affecting your bottom line,” he said.
Cutting down the number of round trips to the server in order for Web pages to load is key for Web content optimization — and in gaining more page views. Stingray 9.0 enhanced with Aptimizer can optimize Web content up to four times faster, while eating up 40-60% less bandwidth, Prabhu claimed, noting that one unnamed corporate website gained 10 more page views per day after deploying Stingray 9.0.
Depending on the enterprises’ environment, the latest version of Stingray — a software-based ADC — can be deployed standard server hardware on Cloud Steelhead to optimize any Web or cloud-based applications for the customer or cloud provider.
ADC vendors are starting to notice that while the legacy ADC model is important for bandwidth optimization, the user’s web experience must come first.
“Users don’t want to be sitting in Starbucks waiting for Web pages to load,” Prabhu said.
Today starts the new era of the Internet — an era in which “the world” comes together to launch the latest Internet protocol, IPv6. It is because the IP address space of it’s primaeval counterpart — IPv4 — is really running out. This limitation must force all, who are interested in providing content to the online community, to move to IPv6 and learn to love it.
The last World IPv6 Day (June 8, 2011) served as a test for IPv6 — to turn it on and find out where the kinks were; this year, organizations who participate in World IPv6 Launch Day — like Google and Facebook — have turned on IPv6 and plan to leave it on permanently.
“World IPv6 Launch Day is a lot larger than people understand. IPv6 is the single largest upgrade in the history of the Internet. It’s not a small decision for the major content providers to turn on IPv6 and leave it on. Going forward, Internet users will be forced to go through transition gateways to reach businesses that do not similarly upgrade by adding IPv6 connectivity, with the result being slower connections and services for their customers,” said John Curran, CEO and president of the American Registry of Internet Numbers (ARIN).
However, last year’s test bodes well for today. ARIN waited by the phones for issues to arise during IPv6 tests, but all they heard were crickets, Curran told me during this video interview about World IPv6 Launch Day.
“The good news is that, done properly, Internet users globally will not notice any significant difference in their daily lives,” he said.
In order to transition to IPv6 properly, enterprises need to have the right tools. On SearchEnterpriseWAN.com, we gathered resources for today to help organizations turn on IPv6: