WAN Virtualization

The RAID Analogy

by Talari Blogger-in-Chief Andy Gottlieb...

When I’m asked for the “elevator pitch” on what Adaptive Private Networking is, I always go back to the RAID analogy – “APN is doing for WANs what RAID did for storage”.  Since we’re not on an elevator, let’s examine this analogy in some detail.

The analogy works both at the business benefits level, and at the technical level.  We’ll look primarily at the benefits part in this post, and the technical parallels in a future post.

Before RAID, in the 1970s and the 1980s there was a large market for expensive mainframe attached storage systems, which IBM dominated, and similar markets for minicomputer storage.  IBM’s margins were extremely high, and even the vendors of “plug compatible” storage systems had high gross margins as well.  They competed on the benefits of their “proprietary disk drive technology”, and only within a narrow range (a factor of 2 or 2 ½ at most) on cost.

Along came Seagate hard disk technology, targeted at the nascent but fast growing market for personal computers. Each hard disk was not nearly as reliable, having nowhere near the MTBF or seek times, of the mainframe/mini disks, but were built in high volumes with massively better price/bit.

RAID - Redundant Array of Inexpensive Disks took advantage of that Seagate disk price/bit to revolutionize the storage market. RAID solved the reliability and capacity limitations of the PC hard disk, and enabled them to be leveraged for data center business quality storage systems.  With RAID technology, companies like EMC built bigger, faster, cheaper and more reliable storage subsystems.  Within a very short time, no one tried to build big, fast reliable proprietary single disk storage systems any longer.

For APN, the existing high margin market is that for private WANs, which today means MPLS and Frame Relay.  Where IBM was the single dominant storage vendor before RAID, today with private WANs, the oligopoly of AT&T and Verizon (and to a much smaller extent, Sprint) in the United States, and government-sponsored telecom monopolies or near monopolies in other countries dominate the market, with 1 or 2 vendors in each company controlling 70% or more market share for these services.

Pricing of private WAN service, which was somewhat aggressive and clearly related to the cost of the service when Frame Relay was introduced in the early 1990s, now bears no correlation whatsoever to the cost of providing the service.  Rather, due to the last mile access (copper and fiber) monopoly by the RBOCs and the oligopoly of nationwide / global private WAN service providers, the price / bit of these services has barely come down in the last 10 years, even as the price of almost everything else has followed the classic Moore’s Law curve downward.

The irony here is that the analogous enabling technology for APN is the public Internet – much of it provided by these very same service providers!  Just as with the PC and the PC hard disk in the 1980s, massive investment in the Internet and broadband has been targeted at individuals and small business, and given the combination of high volumes and competition in that market, the price/bit here has continued to come down with Moore’s Law.  Internet access now costs $3 - $15 / Mbps per month, versus the $450 - $1500 / Mbps per month which the telecom SPs charge for MPLS.

While Talari’s APN technology is pretty sophisticated in terms of how we do real-time per packet measurements and sub-second response – more on that next time – just as with RAID, the price/performance comes from providing the surrounding reliability to enable leverage of Internet economics.  Where RAID wrapped a layer of hardware and intelligent software around multiple PC hard disks, APN’s two-ended, appliance-based solution does something similar with multiple WAN connections - existing private WANs and high speed Internet connections (T3, OC3, Metro Ethernet, etc.) at data centers and large sites, as well as existing connections and any type of broadband Internet links for branch/smaller locations.

RAID leveraged the PC hard disk technology to revolutionize business storage cost, capacity and reliability.  APN leverages the most powerful, ubiquitous, low cost communications network ever created – the public Internet – to deliver Enterprise WANs that are far higher bandwidth, far lower cost and more reliable than the best proprietary single service provider WANs available today.

It may well take a bit longer for APN to change the networking world as completely as RAID changed the storage world, but the parallels in terms of surrounding superior economics with business-class reliability are striking.

What would you do with 20 times the branch bandwidth?

by Talari Blogger-in-Chief Andy Gottlieb...

Fish don’t notice water.  It’s just a given of their environment.

Similarly, scarcity of enterprise WAN bandwidth is a "fact" that IT managers have long taken as a given.  Despite 10 Gbps LAN backbone and server links today, and a minimum of a switched 100 Mbps Fast Ethernet link to every desktop, with Gigabit Ethernet desktop connections more and more frequent, network managers – and application developers, too! – live with the reality that a 1.5 Mbps T1 WAN connection shared across 5 to 50 people is frequently all that’s available.  Even at larger sites, average WAN bandwidth per person will often be similar (or even lower). 

This 3 or 4 order-of-magnitude difference in the availability of bandwidth on the WAN has major implications for network design, server deployment and applications which can be run on the corporate WAN.  And in tough economic times like these, spending twice the amount on your WAN to get more bandwidth to run applications is just not an option.  Heck, spending the same amount you did last year is tough for some companies to do.

When we talk to customers about Adaptive Private Networking (APN) and its benefits, some customers pick up on this idea that they can now afford to have 20 or 30 times the WAN bandwidth they’ve been forced to live with, but for many, it’s an idea that’s so foreign they’ve not though much about the possibilities.

A great example of an application APN enables is videoconferencing.  “Gee, I’d never run ‘real’ videoconferences over the public Internet” is what any network manager worth his salt knows, after all.  That’s what QoS on MPLS is for. 

At Interop Las Vegas 2009, we’ll be demonstrating how having a lot of bandwidth changes the way networks are designed and applications are run.  While almost all techniques from WAN Optimization vendors like Riverbed and Blue Coat are about conserving bandwidth, APN actually takes the opposite approach, focusing, much as RAID did with “cheap” PC hard disks for storage, on making inexpensive bits business quality. 

Because of the inexpensive nature of Internet bandwidth, not only does APN make it affordable to have enough bandwidth to run high definition videoconferencing to the smallest of branch offices simultaneously on your now “converged” network with the rest of your data traffic, it actually duplicates videoconferencing and VoIP sessions to enable very high quality even over “pretty good but by no means perfect” public Internet connections. [Note: such duplication is administrator configurable, and can also be done dependent on the availability of bandwidth.]

So if you’re in Las Vegas next week, please stop by booth 1567 and see for yourself the video quality APN enables even in the face of network conditions which would ruin a simple voice call.

And start thinking about the applications you could run, and the centralization redesigns which would be possible, if only you had 20 times the bandwidth on your WAN. 

Because now you can afford to.

Why is this the right time for an Enterprise WAN revolution?

by Talari Blogger-in-Chief Andy Gottlieb...

Welcome.

Talari is doing for Enterprise Wide Area Networks what RAID did for storage – delivering a network with 30 – 100x bandwidth/$, monthly WAN costs reduced by 40% - 90%, and with greater reliability than existing private WANs using Frame Relay or MPLS.

In a carrier pricing environment where a price/bit factor of 2x is enormous, Adaptive Private Networking brings Moore’s Law and Internet economics to Enterprise WAN buyers for the first time in 15 years.

To answer the question in the title, I’d like to start by telling you a story.

It's a story about the market for mainframe computers, with obscenely high margins and dominated by a single vendor, and later minicomputers with a handful of vendors and very high margins as well. Along came the inexpensive personal computer, and not much happened to the proprietary mainframe and minicomputer business – until those “consumer” PCs were connected by LANs, interconnected by WANs, augmented by client-server computing and reliable distributed software – and the PC hardware / software and LAN/WAN industries grew a huge market, built an ecosystem, and took huge dollars out of the hide of the mainframe and minicomputer budgets.

Wait – that story has already played out…

Ok, let’s talk about the market for expensive mainframe attached storage systems, “plug compatible” with their own extremely high margins and “proprietary disk drive technology” and dominated by a single vendor (IBM). Along came Seagate PC hard disk technology. Not much happened at first. But then along came RAID - Redundant Array of Inexpensive Disks. With RAID technology, companies like EMC built bigger, faster, cheaper and more reliable storage subsystems, and built multi-billion dollar businesses leveraging this “consumer” hard disk technology, and grew a huge market and ecosystem out of the hide of the proprietary mainframe and minicomputer storage budgets. Today, no one tries to build big, fast reliable proprietary single disk storage systems any longer.

Hold on – that story has played out, too…

Actually, this is a story about corporate WANs and the extremely high margins of an industry - the telecom SPs - dominated by one or two vendors in each country. And then along came the Internet and broadband, and… well, in fact all these years later, nothing much has happened to the base corporate WAN business. Sure, Internet connectivity and broadband (cable, DSL) have become a large market, first for surfing the web and email, then e-commerce, etc. Internet broadband is actually the growth business for the telecom SPs, and they continue to invest heavily in these networks, focusing on consumer adoption rather than on enterprise needs.

VPN technology came along and business began to use the Internet for remote access, and in some cases for backup connectivity. But private Frame Relay and MPLS WANs, with per megabit prices of $450 - $1500 per month, remain the dominant way to build corporate Intranets, and are just about the last holdouts from the pressures of Moore’s Law and Internet economics. They are a $25B annual business, including $15B+ for higher bandwidth (fract-T1 and above) connectivity.

VPNs make the Internet secure, and Internet and broadband performance and reliability have gotten a lot better over the last several years. But in fact Internet connectivity on its own does not meet the 3½ nines (99.95%), let alone four nines (99.99%), standard of reliability that businesses have come to expect of their Enterprise WANs. And so even though there is now literally a 100x difference in the price/bit of broadband – at $3 - $15 / Mbps – versus the price of private WAN links, very few businesses, especially larger businesses, have moved off of those expensive, proprietary networks they’ve been using for years, despite the fact that pricing has stayed so high, and bandwidth at most branch offices is less than what’s available to the average home broadband user.

Enter Talari Networks and Adaptive Private Networking (APN).

Just as RAID wrapped a layer of hardware and intelligent software around Seagate hard disk technology, APN does something similar with multiple WAN connections - existing private WANs as well as any kind of Internet WAN links, be they DSL, cable, fiber, Metro Ethernet, WiFi/WiMax, etc.

APN technology uses end-end algorithms to do dynamic, real-time, per-packet traffic engineering.

Rather than optimize each application for the network, APN optimizes the network fabric for all applications.

 APN enables businesses to build a WAN that is simultaneously less expensive, much higher bandwidth, much lower cost and with lower ongoing operational costs than today’s proprietary, single vendor WANs.
APN solves the reliability problem with running VoIP safely on your IP WAN, and the cost and reliability issues with running videoconferencing on your IP WAN.

 And APN can be added seamlessly into existing networks, rather than requiring a forklift upgrade per site, as migration between services even from one’s existing carrier can require.

So APN unlocks the stranglehold that AT&T and Verizon have in this country on their customers’ WANs and WAN budgets. And it does this by leveraging those same carriers (and the cable folks, too) “consumer” broadband networks.

The current WAN Optimization industry, led by Riverbed, is built on the long-standing assumption that your existing network is as reliable as you need it to be, but bits are expensive – and the need, therefore, is to squeeze as much out of those thin pipes as you can.

The WAN Optimization companies are focused on using hard disks to do compression, and the issues of getting bits onto and off of hard disks scalably and at high enough rates. They’re also focused on doing application-specific optimizations to deal with the fact that there is so little available bandwidth on today’s corporate WANs, and on selling the soft benefits of “application acceleration”, since they quickly learned that the 2x to maybe 4x price/bit benefit they deliver over existing networks isn’t a compelling enough reason alone for most customers to deploy it everywhere.

Talari’s insight is that in fact, there are lots of cheap bits out there - they just need to be made reliable enough to be business quality.

A 2x change rarely moves markets, but an order-of-magnitude-plus change does.

With Adaptive Private Networking, Talari intends to build a very big business – for ourselves and our ecosystem partners – out of the hide of that $15B+ Enterprise WAN service budget, by enabling customers to attain huge savings on one of the very biggest line items in their monthly IT budget, 2 orders of magnitude bandwidth per dollar, and equal and in fact better reliability and application predictability than their current locked-in WAN delivers.

More – a lot more – to come.