Blog

Applications don't interfere with cloud adoption... it's the network!

Last week on TechTarget's Cloud Computing web site, Carl Brooks wrote a piece where he interviewed some IT people in the financial services and other IT-hungry industries about their use of Cloud Computing.  His and their conclusion is that cloud computing adoption will be limited by the performance that critical applications can expect when connecting to remote cloud-based resources.

In the article, he discusses other issues such as the quantity of applications in a typical enterprise IT organization and the infrequency with which older applications get retired.  The migration of applications that were originally written for local environments to cloud environments is another area of discussion. 

Obviously application performance is a major success factor in a pretty much any major IT project, including cloud migration. While Riverbed can't help with migration or reduction of applications, we can most definitely help with issues around application performance when those applications are deployed across a WAN. 

If users are in one or more locations remote to the data center where applications are deployed, then it's a call for Riverbed Steelhead appliances.  We used to call that architecture distributed computing, but nowadays, it's called a Private Cloud.  If you have physical access to both sides of the connection, then Steelhead appliances are the way to go.

If you have applications and other services within Amazon Web Services’ Elastic Cloud Compute (EC2) and Virtual Private Cloud (VPC) environments, then you can get the same kind of performance improvement with Cloud Steelhead.  Since you can't physically enter Amazon's data center and install a Steelhead appliance, a different approach is required, and that's what Cloud Steelhead gives you. 

And if you've got a hybrid cloud (which is really the most common arrangement, after all), you'll need a hybrid solution.  But the bottom line is that if you've got distant computing resources, private or public, then you need WAN Optimization to make sure that your users get the kind of performance that they need to be successful and happy.

Regardless of the architecture, Riverbed will make sure you get the most out of your remote applications.

Taneja Group asks, "Is cloud-enabled DR is ready for prime time?" Riverbed says YES!

(by Steve Riley and Evan Marcus)

Last month, in an article at InfoStor.com, Jeff Byrne and Jeff Boles of The Taneja Group wondered aloud, Is cloud-enabled DR ready for prime time?  Here at Riverbed, we can confidently answer yes, indeed it is.

Implementing processes and technology for disaster recovery is like going to the dentist: no one enjoys it (well, perhaps dentists and this kid do) but if you don’t do it, eventually your organization will be in a world of hurt and you might be out of a job. Jeff and Jeff quote an informal statistic that 25% of their readers have experienced unplanned outages. We hope you’ll never be part of that. If, however, someday happens to you, then with a little planning, your trip to the DR dentist will go much easier.

We agree with the claim that cloud-based DR eliminates much of the pain traditional approaches impose. While the authors correctly state that cloud-based DR is “a recipe matched to SME needs,” large enterprises, too, can derive the same benefits. In their article, Jeff and Jeff enumerate six challenges that cloud-based DR brings. We’ll briefly discuss how Riverbed’s new Whitewater Cloud Storage Accelerator Appliance (or just "Whitewater") addresses each of them.

1. Ease of Use: One of the biggest challenges that IT organizations face as they migrate DR to the cloud is that they have to change their infrastructure. They need backup utilities that speak cloud protocols. And then they still have to maintain the old stuff so that they can recover data from the existing DR infrastructure. With Whitewater you don’t have to change your existing backup product at all. You just point your current backup tool at the Whitewater’s NFS or CIFS front-end, and send the backups there just as you always have. 100% compatibility. No need to change technologies. What could be easier?
2. Security: Whitewater protects your data in transit with SSL and in storage (local and cloud) with 256-bit AES encryption. The data is never stored or transported in cleartext after it leaves its original location. The encryption key is stored in the Whitewater appliance and if something happens to the appliance, you just need to enter the key into a new appliance. And only you have the key. Not Riverbed and not the cloud provider.
3. Access: Whitewater does something very unusual in the way it handles data that has passed through it into the cloud. It keeps a local copy of the data (encrypted, of course) so that the data can be accessed locally without touching the remote cloud. The result is dramatic improvements in speed and performance when restoring cloud-based backup data. Whitewater also uses the local data store to deduplicate data that’s going out onto the cloud, reducing your cloud costs.
4. Recovery Time: Since a copy of the most recently backed up data is stored locally within Whitewater, there’s no need to perform most data restores across the cloud. When data does need to be pulled in from across the cloud, it is pulled into Whitewater deduplicated, vastly reducing the time and amount of data that must be moved. The data is reconstituted inside Whitewater and returned to the requesting application in its original form. Apart from the decrease in recovery time, users can’t tell that anything has happened to their data.
5. Recovery Effort: Administrators don’t need to change the tools that they use to recover data through a Whitewater appliance. They can use the same backup and restore utilities that they use today. The Whitewater appliance will translate between the backup protocol and the cloud protocol and retrieve the backed up data with no changes to procedures and no learning curve.
6. Provider Lock-In: Whitewater supports two cloud storage APIs today: Amazon S3 and EMC Atmos. You can easily move from one cloud to another simply by changing Whitewater’s destination.

When evaluating cloud-based DR, a question equally important to “how do I get my stuff backed up?” is “how do I get my stuff back after an outage?” This is where Whitewater really shines because of the constant movement and near-instant recovery capabilities of the cloud-via-DR-gateway architecture. How so? The authors allude to it at the end of their article:

Emerging cloud gateway vendors, when enabled by primary storage support and the ability to serve up storage from a virtual appliance, can in effect pre-stage data automatically and make it available to virtual servers in the cloud. Not all cloud gateway solutions can do this, but when they can they will further simplify the recovery process.

Whitewater can do this today. Let’s walk through a hypothetical.

Six months ago you deployed a Whitewater appliance on your network, pointed it to an Amazon S3 bucket, and retargeted your backup software to the appliance. All of your data is securely backed up offsite. Now the Bad Day happens: your phone rings incessantly, bloggers are griping about you, people paid more than you are mobbing your office. Now what? Well, since your data is already in Amazon Web Services, perhaps the AWS cloud can become your temporary data center? With Whitewater Virtual Edition you can start a compute instance that runs a virtualized appliance, import your encryption key, and obtain access to your data. Then you can start some application compute instances, extract your backed up data into live disk volumes, remap your DNS entries, and get back online.

No, it isn’t exactly point-and-click, at least not yet. But with some planning—mostly defining your compute instance types, understanding scale-out requirements, and testing application behavior—you can bring mission critical (read: revenue generating) applications into service in a matter of a few hours. With that out of the way, then, you can begin work on returning to normal.

There…that wasn’t so bad now, was it? Polish your smile—learn more about Riverbed’s cloud-based disaster recovery solutions today.

How VDI will break service provider offerings

I had a fascinating presentation and discussion with a Riverbed customer yesterday. He’s the cloud computing architect for a global manufacturing company, with about 40 locations worldwide. One fact of note is that their entire organization runs on VDI – that’s impressive in and of itself. But what was more interesting to me is that he believes all service providers are currently at least 5 years behind in enabling him to run VDI the way he wants to. In short – he believes service providers are broken when it comes to the next wave of distributed computing.

The architect used a couple of great parallelisms that couldn’t help but make the topic stick in your mind. First – look at the last wave of traffic to move onto the wire – voice. In many cases it was massively challenging for organizations to engineer their environments to support voice effectively throughout – especially if you were including very remote sites or locations that used public internet or satellite as their connection. In the beginning he argued much of this happened through organizations engineering around service providers, instead of with service providers. (Full disclaimer – these are the claims of the customer, not of Riverbed or myself!)

At the time, voice was the essential communication mechanism that could not fail. People frequently spent hours a day on the phone, so both quality and reliability were essential. But another interesting phenomenon happened along the way – the cell phone. Somehow the world at large was OK with going backwards in quality and reliability (does the phrase “Can you hear me now?” sound familiar?). For the enterprise, why was that acceptable? The customer posited two reasons – for one, the benefit of mobility was a productivity enhancer, but two, the phone was simply no longer as important as your interaction with your computer.

For a knowledge worker like myself, this couldn’t be any truer. On a busy day I might spend 1-2 hours on the phone. The rest of the time I’m probably doing something with my computer. In fact, in many cases when I’m on the phone I’m on the computer too! I could much more easily survive without telephone communications than I could survive without computer access in general.

And now back to VDI. For our customer, he discovered two major problems with regards to how he uses his network today. His service provider has difficulty separating out “real-time traffic” like voice from “interactive traffic” like VDI, where intense back-and-forth WAN communications could massively limit the effectiveness of a worker. Moreover, it’s not just downtime that would affect him, but even high latency would drive losses of a few hundred thousand dollars per hour, per office. And since his service provider only reports average latency for the past month, he’s out of luck in terms of proactively addressing the problem.

The other challenge he has is, in some way, to model the way transitioning to VDI or making changes to your setup will impact your end user experience. For example, if you switch to Windows 7 virtual desktops it’s likely that the bandwidth required per session will go up about 40%. It’s easy to say how much network capacity you’ll require based on number of users, but how will a blip in latency impact the user experience? Will it change on Windows 7 versus other platforms?

Technology is going in the right direction – though the architect needs to see it move faster. For example, he can use Steelhead appliances today in order to prioritize VDI traffic above everything else, even voice and video. But the ability to effectively and scalably model the user experience of VDI is a hurdle that he has not been able to overcome, beyond simply making his users or staff sit at a desk, and spend time testing software instead of closing business.

I would be interested in hearing from others – do you have the tools that you need to implement VDI across your distributed environment? 

 

On the proper use of naughty words around the Riverbed office…

The “S” word has been thrown around a lot lately at Riverbed HQ in San Francisco, to point where our executive team thought it was time to do something about it.  No, not that “S” word, this is a family-friendly place after all.  The “S” word I’m referring to is solutions as in the sentence: “Wow, them sweet Riverbed solutions sure have saved my bacon!”

The problem with this particular “S” word is it gets bandied about so often it’s lost all meaning, much like that other overused attention-getter the “C” word, by which I mean cloud of course.  The word solution is trotted out there any time someone wants to sound like they’re selling something more than a point product.  The word cloud is tacked on anytime a company wants to increase its stock multiple. 

Right, then if these words are bad, why are we using them so much?  It all goes back to the basic reasons these words got so popular in the first place.  A solution by definition solves problems.  The cloud by definition introduces new options on how to do IT.

Riverbed is having a serious re-think as we evolve from a single technology offering into a grown up IT-performance player.  As phenomenally cool as Steelhead appliances are, we’ve decided it’s high time we broadened the portfolio.  Why?  Simple answer: to better serve our customers needs.

So rather than being just an extremely profitable WAN optimization box shifter, we bring to the table proper solutions that include:

• Application Aware Network Performance Management for more visibility, faster problem resolution, and better alignment of IT to the business needs
• Cloud Storage Acceleration for easy integration of data protection with public cloud storage and the promised benefits
• Advanced Quality of Service to better control and prioritize traffic, not just make it faster and slimmer
• Vertical Expertise to address the biggest pain points of your specific industry
• Professional Services for training, consulting, and award winning support around the globe
• Technology Alliances with Industry Leaders for advanced features and reduced integration risk
• A wide range of Channel Partners for added value in bundling and delivering complementary offerings, giving you more buying options

What do you get when a company broadens its focus beyond just having the best flagship WAN optimization product family?  You get well rounded, customer-oriented solutions to solve real challenges you face in IT performance.  That, my friends, is a solution, and that’s the new Riverbed vision. 

Delivering Local-like performance for RemoteFX over the WAN

Today, Microsoft announced the availability of SP1 for Windows 2008 R2.  One of the key capabilities included in SP1 is RemoteFX, which is Microsoft's next-generation desktop virtualization platform. 

RemoteFX leverages the power of virtualized graphics resources and advanced codecs to recreate the fidelity of hardware-assisted graphics acceleration, including support for 3D content and Windows Aero, on a remote user’s device. This allows for a local-like, remote user experience over a variety of possible host, client and network configurations. Dynamic Memory is a new memory management feature of Windows Server 2008 R2 Hyper-V. 

RemoteFX is optimized to deliver performance to users located on a local area network.  Riverbed extends the local performance experience to RemoteFX users located over a WAN.  Riverbed's Steelhead family of WAN optimization products allow the RemoteFX protocol to more fully utilize the available WAN bandwidth, and significantly increase the number of frames per second of a RemoteFX session, resulting in a smoother, higher fidelity user experience.

Riverbed optimizes Microsoft RemoteFX traffic over the WAN:

• For office-type users over the WAN
• Reduce TCP bandwidth consumption over the WAN, increasing productivity
• Supports remotely rendered WMP files, and streamed WMP, Microsoft Silverlight, and Flash video 

Here is a demo of Riverbed accelerating RemoteFX

 

You can find out more about the RemoteFX SP1 announcement today by reading Microsoft's blog. You can also download a trial of of RemoteFX SP1 here.

 

 

Another glaring example of the confusion surrounding cloud storage

Tom trainer recently blogged in Network Computing about how EMC could become the Amdahl of cloud storage.  Tom starts off his blog by saying that there are some fundamental issues around how EMC sells its Atmos cloud storage platform to customers.  The key issue he focuses on is how EMC charges for storage up front and not for storage that is used.  According to Tom, the utility aspect of cloud storage is missing from Atmos.

This is totally incorrect as EMC sells Atmos to service providers such as AT&T, who in turn sell cloud storage to their customers.  AT&T's service is called Synaptic and it is indeed based on a utility or pay for what you use model.

Personally I believe the more accurate cloud storage discussions should be around how to address the roadblocks that exist when deploying cloud storage.  Specifically, how to deploy cloud storage without changing your backup platform ecosystem.  What about ensuring security when transferring data to the cloud and ultimately storing your sensitive data in the cloud?  And obviously performance?  If you move to a cloud storage environment, the performance when accessing that disk in the sky will be much slower compared to your local area network environment. 

I will end my rant with a shameless plug.  Riverbed's recently launched Whitewater product addresses the cloud storage roadblocks and enables organizations to seamlessly deploy cloud storage and reap all the cloudy benefits that you would expect.  This includes paying for only what you use.

You can watch a demo of Whitewater here:

 

 

 

 

Space – the final frontier (for WAN optimization)

Today's guest blogger is Naveen Prabhu, Product Marketing Manager here at Riverbed.

In November of last year, we acquired a company called Global Protocols, the leading developer of commercial SCPS-TP. SCPS-TP or the Space Communications Protocol Specification – Transport Protocol (if you are not into brevity) is a protocol that enhances data communication over high latency, low bandwidth links.  These characteristics define the problems encountered by users of Satellite Networks. Such networks also tend to be affected by a higher degree of loss (Bit Error Rate), and this makes standard TCP very inefficient (think: slow start, smaller window sizes).

Skipware, as Global Protocols' implementation of SCPS-TP is known, has become, over the years, the de facto standard for protocol optimization over satellite networks in Government and DoD deployments. Skipware can do a lot of things regular WAN optimization can't , and conversely WAN optimization can offer certain things like data deduplication that Skipware does not support. 

It is this synergistic combination of the two technologies that makes 2 + 2 = 5.

Riverbed's Steelhead appliances offer native support for LTTS (Loss Tolerant Transport for Satellite) as well as TCP Westwood, which provides satellite optimization bandwidth estimation (yeah, we need to come up with a better marketing term for this one!), Skipware acts as a catalyst to these native improvements and gives an extra boost to data communication over these skinny high latency links.

So whether you're working out in the field on a top secret mission or vacationing on a cruise ship, you can be rest assured that Riverbed's got your back!

Sharing some Riverbed love

I've written before about how great it is to be part of a company about which customers, columnists, and industry analysts consistently say such nice things.  I thought that for Valentine's Day I'd share some examples of the love that Riverbed gets shown on a regular basis. 

One of the places that our users get to show us some love is in our Riverbed Performance Hall of Fame (PHOF).  We encourage our users to brag about the performance improvements that they've seen since they implemented their Steelhead appliances with text and with graphics.  Here are a couple of examples with links back to the listings on the Performance Hall of Fame page.  You'll find lots more on the page itself.  (When you visit, please be sure you spend a few minutes awarding Kudos (votes) to some of the entries that you like.)

Steelheads Rev Up Internet Cafes in Afghanistan.  My recent favorite comes from our troops in Afghanistan, where they use Steelhead appliances to enhance performance at Internet Cafes that support the Morale, Welfare, and Recreation (MWR) needs of the troops.

540.84X BW Optimization to one application ... Amazing :) This unbelievable statistic is shown graphically in this PHOF entry.  They saw a reduction in bandwidth of better than 99%, cutting WAN traffic from 1220GB to about 2 1/2 GB.

Analysts are showing us some love too.  The results of Gartner's WAN Optimization Magic Quadrant, which were released last month, made a lot of people smile over here.  And, InfoWorld named us their WAN Optimization product of the Year for the sixth straight time.

We see a lot of love on Twitter too.

So yeah, there's lots of love around for Riverbed.  If you've got wide area networks, you should take a look.  You'll love what we can do for you, too.

10 Important Steelhead Features that you can't get from other vendors

With the recent announcement of RiOS 6.5, Riverbed continues to extend its product and technology leadership.  There are now more features and capabilities that are unique to Riverbed than there ever have been before.  Below is a partial list of the key features available from Riverbed that you will not be able to find in any other WAN optimization vendor's products.  Note that this is a partial list, in that I've omitted unique features such as Lotus Notes and Oracle E-Business Suite optimization that may not be important for many customers.

1)  Optimization for encrypted MAPI:  MAPI traffic is now encrypted by default in Exchange 2007 and 2010.  While other vendors may be able to provide MAPI optimizations, to use those products you have to disable the default encryption setting in the Exchange 2007/2010 server.  But with Riverbed, you get fast encrypted MAPI performance without compromising security.

2)  Latency optimization for Outlook Anywhere (RPC over SSL) traffic:  Outlook Anywhere allows Outlook users to obtain emails without using a VPN.  It provides the IT organization with unprecedented flexibility and options for supporting email users.  And now, email delivered through Outlook Anywhere can be just as fast as when delivered through optimized MAPI when using the Steelhead solution.  Riverbed is the only WAN optimization vendor that offers this capability.

3)  Optimization for SMB-Signed CIFS traffic:  SMB-signing is a Microsoft feature providing for enhanced authentication security for CIFS traffic.  Riverbed is the only vendor who can provide both bandwidth and latency optimizations for CIFS traffic that is protected through SMB-signing.

4)  SMBv2 Optimization:  SMBv2 is the new CIFS protocol available in Windows 2008 R2 servers and Windows Vista and Windows 7 workstations.  While it is somewhat less chatty than the previous SMB version 1 protocol used in earlier Windows software, nevertheless there are some protocol inefficiencies.  Riverbed is the only WAN optimization vendor capable of natively providing both bandwidth and latency optimizations for SMBv2, without forcing the Windows client and server to use the older CIFS protocol.

5)  Optimization for applications requiring client-side SSL certificate authentication:  In some highly-sensitive environments, password-based authentication simply isn't enough.  When this is the case, sensitive applications may require that the client provide its SSL certificate for authentication before allowing access.  Riverbed is the only vendor who can optimize SSL connections when the application server demands certificate-based authentication for all clients attempting to access data.

6)  Optimizations for EMC SRDF and FCIP traffic:  Since the RiOS 6.1 release, the Steelhead has been able to intelligently apply its algorithms to SRDF and FCIP traffic by recognizing application-specific fields and markers within the data flow.  RiOS 6.5 introduced a new feature that selectively applies the appropriate optimization level for data from common RDF groups within the SRDF session.  This new feature conserves storage capacity in the Steelhead, and leads to further improvements in overall data reduction effectiveness.  As with the other items on this list, these SRDF-specific capabilities are only available from Riverbed.

7)  Data Store synchronization:  Riverbed is the only vendor with the capability to synchronize data stores to a redundant Steelhead appliance.  This is particularly important when HA requirements dictate that there be no performance degradation in the event of device failure.  With Riverbed, a Steelhead's data store can be replicated in real time to a clustered Steelhead.  Failure of any Steelhead does not result in a cold data store, because the standby Steelhead has a copy of the failed Steelhead's data store.  Riverbed is the only vendor with this capability.

8)  Mobile Software Client for Macintosh OSX:  A software client is becoming increasingly important as enterprises provide WAN optimization benefits to their mobile users.  With the increasing popularity of Apple products, it's important to realize that Riverbed's Steelhead Mobile is the only WAN optimization software client available to users of the Macintosh OSX operating system.

9)  Optimization-aware dedicated load balancing device:  Riverbed's Interceptor appliance is the only dedicated load balancer device specifically designed to provide optimization-aware deployment scaling.  Specifically, the Interceptor is the only product to perform intelligent load balancing decisions that maximize the probability of a "warm" data store hit when distributing TCP connections to a cluster of WAN optimization appliances.

10)  Cloud Steelhead with transparent discovery agent:  Riverbed offers the only WAN optimization product that is capable of being deployed into a public cloud infrastructure, without the need to install new physical devices or configure switch-based traffic re-routing within that Cloud infrastructure.  This is only possible due to Riverbed's introduction of the cloud discovery agent.

Riverbed supports San Francisco Food Bank

 

In San Francisco alone, 197,000 people struggle each day to feed themselves and their families. 237,000 is the number of people who live at or below 185% of the federal poverty line in San Francisco and Marin - $33,873 per year for a family of three. It's at this income level that children are eligible for Free and Reduced Price Meals and families are eligible for WIC. Very often, these families lack the resources to provide enough food to consistently nourish themselves.

To assist with the effort, San Francisco-based Riverbed Technology recently donated a $25,000 check.  This money is in addition to the ongoing support provided by Riverbed employees as they donate their own money and volunteer their time to help feed the needy.

Here is a clip of Riverbed CEO Jerry Kennelly being interviewed by CBS news.

  

Riverbed Delivering Enhanced Control

Today's guest blogger is Riverbed's Sr. Solutions Marketing Manager for the Steelhead Appliance, Vikram Ramesh. 

Riverbed yesterday announced the next release of the RiOS  v6.5, which powers the company's award-winning WAN optimization solutions and Cascade v9.0, its advanced network and application performance analysis and visibility solution, that provide globally connected enterprises with business-level visibility and control for all types of traffic and all applications.

I wanted to especially focus on the enhancements we have delivered to our advanced QoS engine as a part of this release.

A WAN optimization solution needs to classify applications accurately and apply quality of service rules and policies that help ensure the high performance of business-critical applications. Controlling applications effectively requires visibility to determine which applications need to be protected, which applications should be contained, and how much bandwidth should be allocated to each. 

Most WAN optimization vendors use a combination of deep packet inspection (DPI) and basic QoS schedulers, which fail to enhance applications that are sensitive to latency and packet loss, including business VoIP, video, and interactive Web applications. This combination also does little for recreational applications that port-hop, or that change and maskprotocols.

Riverbed QoS Engine - Delivering Enhanced Control for Fine–tuned IT Performance

With this release of RiOS 6.5, Riverbed delivers the uniquecombination of an application classification engine (AppFlow^®) and an advanced scheduling technique based on Hierarchical Fair Service Curves (HFSC) that not only allocates minimum and maximum bandwidth, but also prioritizes applications based on their latency sensitivity. To simplify and accelerate the setup of QoS policies, Steelhead offers built-in policy templates that deliver plug & play deployment.

Riverbed AppFlow – Application Classification Engine

The Riverbed AppFlow engine utilizes a variety of techniques, often in combination, to maximize the accuracy and efficiency of its real-time network traffic classification. These techniques include port-based classification, application signature matching, protocol dissection, future flow registration, behavioral classification and others, to identify applications that may hop ports or be otherwise hard todetect. The AppFlow engine can identify and classify hundreds of common enterprise applications and is configurable to classify thousands of customapplications. This ensures that critical applications like web, voice and video are protected while recreational applications are contained.

Riverbed Hierarchical Fair Service Curve (HFSC) Scheduling

The Riverbed HSFC scheduler not only addresses the allocation of minimum and maximum bandwidth (as many products do), but also prioritizes and schedules applications based on their latency sensitivity, thus eliminating jitter and starvation of applications. Steelhead appliances can apply these state-of-the-art techniques while remaining compatible with QoS enforcement on routers and other devices. 

In conclusion, the Riverbed QoS engine delivers the highest level of granularity, performance, and ease of management. With Riverbed appliances, customers get best-of-breed WAN Optimization solutions that deliver fine-grained classification and control over the applications that traverse the network,helping deliver the utmost control and acceleration, thereby enabling organizations to work better and faster.

Riverbed announces RiOS 6.5 and Cascade 9.0

Earlier today, Riverbed released updates to our RiOS and Cascade products.  You can find the press release here. 

RiOS is the Riverbed Optimization System.  I think of it as the Operating System that runs Steelhead Appliances.  But this time around, I want to talk about Cascade and its major improvements first.

With this release of Cascade, Riverbed will provide a wide range of functionality that can discover the critical services on your network, prioritize critical traffic with advanced, built-in QoS capabilities, and monitor performance of your applications with easy-to-use, executive-level dashboards, all while reducing your hardware and software costs. This helps to extend visibility from the network engineer all the way to the C-suite, and the ability for IT to control and allocate resources based on the business needs of enterprises.

Riverbed Cascade® 9.0 delivers a set of service dashboards that report on application and network status in an easy-to-use format, so you can see how the network is delivering your critical applications. With Cascade dashboards and reports, you can finally answer business and technical questions about your network around what are the performance problems, where did they occur in the service delivery path, identify the root cause, fix it and ensure that your service-level objectives are met.

Some of the new features in Cascade 9.0 are,

• Executive-level Dashboards
• SLA Reporting
• N-tier Application Discovery Wizard

Executive-level Dashboards are designed to give a C-level person an easy way to quickly see the status of all of his managed networks.  When you see a red circle, as in the example to the left, you can drill down and see what the source of the problem is (as is shown under San Francisco).  Cascade 9.0 enables network managers to drill down further, and identify the source of the problem.  You can find exactly where the problem occurs in the delivery path, based on discovery and dependency mapping. 

And unlike competitive products, you can navigate between apps, servers, and users to really identify and solve the problem.  You are not limited just to packets.  You can find how long the problem has existed, whether or not it's getting worse, and even identify the root cause.

The Service Performance Report (right) gives users and administrators a high-level summary view of what's going on across the network.  We've also added QoS reporting per application and network segment, along with real-time monitoring and historical analysis.

These QoS-centric features increase the gap between Riverbed's offerings and our competition.

The new release of RiOS 6.5 adds some major new features as well.  One is the new Riverbed AppFlow Classification engine.  Essentially, it will give users the ability to classify network traffic by the application that sends it.  RiOS will ship with the ability to recognize the most popular and commonly found applications, and users will have the ability to add their own apps, even custom ones. 

And then once the apps have been identified, you'll be able to schedule or prioritize them by QoS levels, and QoS will also be smart enough to take into account applications' latency requirements and sensitivity. 

We're also adding improvements to satellite optimizations, client side SSL certificate support, and support for IPv6, and several industry firsts:

• Native Optimizations for SMB v2 traffic
• Full-spectrum Optimization for Outlook Anywhere
• End-to-end Client server signing of traffic
• Selective Optimizations for SRDF

As always, Riverbed customers with support contracts can get these upgrades for no additional charge.  Watch for other blog entries today with demos and information about these important releases.

Riverbed Connect Podcast - RiOS 6.5 and Cascade 9.0

Bob Gilbert and Vikram Ramesh discuss the 6.5 update to RiOS, the software that powers Riverbed's WAN optimization products.  Also discussed is Cascade 9.0, which is a major update to Riverbed's solution for providing visibility into network and application performance.

 

 

Riverbed Think Fast Demo - Cascade 9.0

Cascade 9.0 Demo

 

Riverbed Think Fast Demo - Advanced QoS Engine

RiOS 6.5 and the Advanced QoS Engine

 

Don't Make Important Choices in a Vacuum!

Let's face it, choosing a WAN optimization product from the various vendor offerings can be confusing and difficult, because all of the vendors seem to say the same things about their products.  They all make similar-sounding claims about fast LAN-like performance, server consolidation, cost savings, etc.  And if what they all were saying were true, then it would be rather a simple matter to just choose a vendor based on lowest price, since supposedly every vendor's product pretty much does the same thing.

And while that is precisely what many vendors would have you believe, the truth is that there are very significant differences among the various vendor offerings in the WAN optimization market.  Tragically, numerous customers have found that out the hard way, making the mistake of purchasing a WAN optimization product that didn't work or scale for their network.

A lab test or live proof-of-concept (POC) may help. A test or POC allows customers to get familiar with the look and feel of each product they are evaluating.  But there are limitations to what a lab test or POC can reveal about a product.  Just about every WAN optimization product will work fine in a protected, limited-scale deployment where they are not fully-exposed to WAN disruptions, heavy data center traffic loads, and other nasty things that inevitably happen in a live production network.  It's a huge mistake to assume that just because a particular vendor product works adequately when deployed to 2 or 3 sites in a POC, that it will also work fine when deployed to 20 or 30 sites.

When considering a large enterprise-scale deployment of WAN optimization devices, the most important step in the evaluation process is talking to another customer with a similar-sized network who has also deployed products from the vendor under consideration.  There simply is no substitute for that.

Checking references involves more than listening to vague anecdotal accounts of "success" by the vendor's sales rep.  Rather, it's very important to obtain the contact info of another customer who has deployed and used the vendor's WAN optimization products, and to have a confidential one-on-one conversation with that other customer.  When talking with that customer, ask about specifics such as the number and types of applications being optimized by the deployed products, number of users and locations in the customer's network,  and overall experience from using that vendor's products.  Make sure that the customer has a similarly complex network as your own, uses the same type of applications, and has deployed a similar number of devices that you anticipate deploying in your own network.

At Riverbed we expect and welcome any such requests for references.  With an installed base of over 12,000 customers, we have numerous Riverbed customers who would be happy to relay their experiences to you.  Unfortunately, that is not true for all vendors in this space.

The January Performance Hall of Fame Winner

We're very pleased to announce the winner of the Performance Hall of Fame for January 2011.  It's Michael Vassallo, Senior Network Administrator with Dancker, Sellew and Douglas (DS&D), founded in 1829, is one of the nation’s leading providers of furniture, laboratory products, and inventory management solutions to the commercial office, educational, and healthcare marketplaces.  His entry, Go BRANCH WARMING!!! 2.3TB Pulled from Steelhead by 1 client alone!!!!!, was the monthly winner.

Michael is responsible for DS&D's LAN, WAN, servers, storage, desktop management, and phone issues.  In his spare time, he does a little web site and database coding, and occasionally eats and breathes.

DS&D has five Steelhead Appliances (two 1050Ms, two 520s, and a 200) and one Steelhead Mobile Controller spread between corporate, their DR site, and their remote offices.  Their top traffic types are AppAssure replication and Equallogic SAN replication, with a bunch of AutoCAD, PDFs, MS Office, MAPI and CIFS as well.  There's a lot of DR data going across the WAN, with some regular user stuff behind it.

When I asked Michael for a story about how his users have benefited from having Riverbed in place, he said,

I could point all the obvious benefits in various stories about cost savings etc., etc., yada, yada.  But everyone hears those.  The best story I can point is what happens when a User who has become accustomed to the optimization no longer has it.  They scream loudly about how SLOW everything is and breathe a SIGH of relief when things are normal again.

He also wants me to mention that if users lose optimization, it's because of network issues that are unrelated to the Steelhead Appliances.  :) 

Michael has been written up as a happy Riverbed customer before.  And we have a video of Michael talking about DS&D's successes with Riverbed.

 

Congratulations, Michael, and thank you for your enthusiastic support of Riverbed.

Competitive Advantage in the Cloud

A recent Harvard Business Review blog authored by Sinan Aral, Arun Sundararajan, and Mingdi Xin discusses the results from research they conducted on cloud computing and what the strategic implications were for firms considering cloud-based solutions.

Although the research sample was relatively small (about 2 dozen CIOs and senior IT managers), the research involved highly qualitative, in depth interviews.  The survey results found that although the cloud promises to create value on economic, technical and strategic fronts, firms report a wide range of performance gains from their adoption. Importantly, and not surprisingly, the firms that orchestrate a set of complementary capabilities report higher returns from their cloud adoption.

The research also found that agility is a plus when adopting cloud services.  Firms that are structured to quickly increase or decrease their commitment to new applications and innovations were better suited to cloud-based solutions, which themselves allow rapid scaling of resources and thus lower the risk of organizational innovation.

Finally, the blog states that their findings suggest that cloud-based models offer advantages for some applications and to some clients but not others.  They are extending the research on this front and posing a survey to hear what your company's experience has been with cloud services.

Riverbed's position continues to be that we focus on the performance aspect of cloud services. Once an organization gets past various cloud computing concerns ranging from security to vendor lock-in, Riverbed will assist with addressing the performance challenges.

 

Kindles, iPads, and future mobile digital content

I am often asked about long-term IT trends affecting Riverbed. Two of the key trends that we track and respond to are the movement toward mobile devices and the movement toward cloud computing.  I often make the point to people that these are related issues, and that it’s unwise to think about them separately: effectively, today’s enterprise clients are lightening up and becoming mobile, while today’s enterprise servers are virtualizing and moving from corporate data centers to cloud provider data centers. 

These trends have prompted us to develop Steelhead variants that can be deployed in more places (like Steelhead Mobile for Mac, Virtual Steelhead, and Cloud Steelhead) while still supporting fully flexible interoperation for optimization.  The net effect for the customer is increased flexibility and agility.  And we fully expect that we will continue to explore new ways to support performance in the evolving IT universe.

But now I’m going to disregard my own advice about considering both the mobile devices and the cloud, and instead share some thoughts about mobile devices and mobile content without making much reference to the servers or services supporting them.  In particular, is it better to have a Kindle or an iPad? 

Having had both for a few weeks, I now think the premise of the question is wrong.  Similarly, when I read commentary from old-school book-lovers about the dangers of e-books, I find myself thinking “no, you just don’t understand.”

A former marketing VP explained the situation nicely when he told me that the press loves conflict. (In fairness to the media, this is mostly a reflection of the fact their readership loves conflict).  It’s an easily grasped story to talk about a battle, who’s ahead, who will win.  So even situations that don’t really have a duel-to-the-death quality get cast into that framework.

The reality, at least as I experience it, is quite different.  First off, Kindles and iPads are in no way substitutable for each other. Yes, you can run reading apps on the iPad.  And yes, there is a half-hearted browser lurking in the “experimental” section of the Kindle.  And, importantly, both are primarily devices for consumption rather than production of content. There’s also the relatively uninteresting point that they are both tablet formats. But that pretty much exhausts the similarities. 

A Kindle is a very interesting take on what published text can be in a digital networked age.  It’s either weak or totally useless for anything that isn’t published text. Meanwhile, the iPad is a very interesting take on how to rethink the full multimedia experience (well… the full multimedia experience except Flash, I guess). It’s OK for text as well, but it’s not a very well-engineered solution if published text is your primary interest.  The Kindle is so much lighter, cheaper, and better designed for page-turning that the iPad looks clunky.  The interesting exceptions are where magazines are being published specifically for the iPad, which is a better match for a certain style of “eye candy” magazines.

So I think the defenders of “old-style” books against e-books and reading devices have their battle lines drawn incorrectly.  When I started writing this item, I was thinking of Jason Epstein’s various pieces in the New York Review of Books, where he has made a number of acute comments about digitization while rarely hiding his sympathies for the codex -- stacked pages between covers -- format.  But re-reading his articles from July 2001, March 2010, and February 2011, I realize that he’s actually much more sophisticated in his analysis than I remembered.  Nevertheless, I do run into people whose basic reaction to digital books is to think that they’re awful.

The Kindle seems like an ally to books – it lets you take books to places you couldn’t have taken them before.  Over Christmas I took a vacation with my family travelling around Vietnam, and during breaks I could read Harold McGee’s magisterial On Food and Cooking.  There is no way I could have taken the 896-page hardcover book with me, nor would I have been able to read it at the breakfast table.  But it was amazing to have not only that book but about a dozen others, as well as subscriptions to the Wall Street Journal and Financial Times, in a tiny package that hardly required any power.  It’s become trivial for me to slip it into my coat pocket or my briefcase, almost no matter where I’m going.  Whenever I travel there’s a high degree of uncertainty about when I will have “dead time” and how long it will be – the Kindle makes it much easier (and easier on my back) to be prepared for long stretches that nevertheless might not be workable for laptop-based activities.

Now, as much as I like the Kindle there are still some notable drawbacks.  The conversion process from book to e-book is not always smooth, and so McGee’s book (for example) has some broken table layouts as well as mostly-useless cross references given in terms of the hardcover page numbers.  The 3G delivery of newspapers is brilliant but Amazon wants to charge me $5 per week to support it outside the US and (adding insult to injury) won’t deliver any graphics even if I’m paying that fee.  It’s cool that I could fall back on using my laptop’s better browser and WiFi to grab an issue of the WSJ and then transfer it to the Kindle via USB – but it wasn’t cool that I had to open up the Kindle’s directory structure and figure out that the file belonged in the “Documents” directory.  (That was another one of those things that was pretty obvious to me but wouldn’t have been at all obvious to the non-computer-focused people in my life).

Returning to fear of e-books, I think it’s more reasonable for a book-lover to be concerned about iPads, because there we have both the current demonstration of new interface models and the tantalizing hints of future hybrid content – the iPad makes it easy to envision a future further convergence of elements of the Web, movies, gaming, and music.  It seems likely that iPad content will be a lot more exciting overall than Kindle content.  But excitement isn’t everything. 

An element of what scholars don’t like about e-books is the threat to their habits and methods.  But from my perspective that battle was lost a long time ago.  I already have the odd experience that when I cite something that’s not available on the web, people tend to be skeptical.  Just 15 years ago it was flaky to use a URL as a citation instead of a published journal or book, now it’s the other way around – because (implicitly) who finds information by a technique other than the Web these days?  And if I can’t immediately follow your link to check it out for myself, who’s to say you didn’t just make it up?  It was very strange when there was a nice article about Riverbed printed only in my local (Boston) version of the Wall Street Journal – not available online, and not printed in the San Francisco edition where the HQ folks would see it.  I wound up clipping and scanning it to share with others.

Everyone seems to understand that an iPad is a rotten substitute for an iPod – why would you want a honkin’ big device if all you cared about was music, and you could get a tiny cheap easier-to-use device instead of the bigger expensive multi-purpose thing?  But for some reason people seem to have more trouble with applying the same reasoning to published text, and thus some of the iPad vs. Kindle nonsense.

 I have been intrigued that the place I really have found an iPad to be valuable is in managing my personal (non-work) email and other tasks, such as visiting social networking sites. I can’t really justify this assessment, it’s more of an observation at this stage.  I suspect that some kind of future MacBook will do better at combining the laptop productivity that I need at work with the touchscreen gestures that I find congenial for non-work… and perhaps one of the Windows tablets could do that for me today.  But I definitely find that I’m doing a better job of keeping up with my personal email since I started handling it mostly via iPad.  

 Of course, it’s possible that the “iPad vs. Kindle” pseudo-contest is really just a reflection of the Apple cult at work… perhaps the real problem with the Kindle for some analysts is that it doesn’t come from Apple. Perhaps we can look forward to a Kindle-like device from Apple (the iRead?  iText? iBook?) before too long.