• Lori and I have a (in good fun) BATTLE ROYAL over the term “ADC” versus load balancers.
• I have single-handedly renamed Direct Server Return (DSR) to “Bat-shit Crazy Mode”.
• We talked about challenges with providing accurate health checks
• We were distracted several times by squirrels.

Have a listen, good stuff.

The fact is, we’re more addicted to cookies than even Cookie Monster

Came upon this post on Slashdot.org, a criticism of HTTP cookies entitled “HTTP cookies, or how not to design a protocol“.    It goes into many of the security issues surrounding HTTP cookies, and how the popular criticisms (privacy) aren’t the real problems.

The problem is, HTTP cookies are absolutely vital to everything we do on the Internet. Any website that we visit where the server builds up customized content (even if it’s just selection which region we’re from, like on Fedex.com), depends up on cookies.

Want to play a nasty prank on a co-worker who doesn’t lock their screen when they leave their desk?  Disable cookies on their browser.  Nothing works.

Cookies are just about the only mechanism in use to create a unique relationship between a client and a server. In other words, a cookie is the only way to establish a session.  Other than cookies (or long URLs), the HTTP protocol does not

In the end, we’re worse than cookie monster with our cookie addiction.

Final purchase price was around $18,000,000.  Not a bad deal for Radware (considering Nortel bought Alteon for $7.8 billion in 2000).

But speculation on how Sun will die or be consumed isn’t the point of this post: The point is that load balancers had a hand, at least indirectly, at killing Sun.

All in the IT world are familiar with what killed Sun:  Commoditized hardware.  Load balancing is what made commoditized hardware viable as an Internet platform.

When you need to add capacity, you can either scale horizontally, or vertically. Vertical scaling means adding more processors, better processors, more RAM, more disk, etc., to an existing system.  Horizontally means adding more systems.

Sun hardware was great at scaling up.  But it was expensive.  Way more expensive.  But initially that’s what people did.   If only there was a way to split the load amongst smaller, cheaper systems, giving them the power of a vertical system at a fraction of the price.

Oh wait.

I started out as a Solaris administrator in the late 90s. In the geek IT community, Sun workstations where the ultimate materialistic status symbol.  They were faster than x86 chips, and Unix had a solid reputation as a platform, with Solaris being the best of the Unix breeds.  Those where the halcyon days.

I remember building out a Sun server farm for a client around 1999, consisting of about 50 or so Sun boxes.  We were happy with the hardware and software, but the prices was something to behold.  Sun’s idea of a web server was a Sun E250, which moderately equipped cost in the $30,000 range.

A $30,000 web server.

But hey, it was the late 90′s.  They would dump garbage trucks worth of cash on Internet infrastructures, so the high price wasn’t that big of a deal.  Windows was only moderately trusted as a server platform, and Linux was still considered a hobby platform, with little cachet outside of its dedicated fan base.  Solaris ruled.

But things changed.  With the dot-com bust of 2000, people (rightly so) choked at the idea of spending $30,000 on a web server.  Linux had grown exponentially, and was trusted for web and application serving.  Windows had grown to be a strong web platform as well. As a result, you could buy an x86 web server running Linux or Windows for about $2,000 that had roughly the same ability to serve traffic as a $30,000 Sun server.

Load balancing made commodity hardware viable as an Internet platform. It didn’t matter much what processing power a web server had, you just threw more behind a load balancer until you had the capacity you needed.   Scaling horizontally was much less expensive than vertically.

Linux exploded in the server realm and Windows grew stronger, all at the cost of Sun.  It still made sense to run your database on vertically scalable Sun servers, but the web and middleware markets were lost to Sun forever.

For too long Sun refused to acknowledge that $30,000 was too much to charge for a web server.  They even dropped Solaris support for x86.  They’ve taken a number of positive steps, such as offering x86 servers in addition to SPARC, and fully supporting Solaris on x86 and even open sourcing Solaris with Open Solaris.  But it seems to be too late.

I was recently at a tier 1 hosting data center, and the racks are full of servers from HP and Dell.  10 years ago, tier 1 data centers where so full of Sun equipment, that all you could see was purple and gray.

I was perusing the latest issue of Inc. magazine, and low and behold, there was an article featuring Selina Lo, the VP of marketing for Alteon when they were purchased by Nortel in 2000 for $8 billion.

Currently she’s the CEO of Ruckus Wireless, a Silicon Valley startup.  The article delves into self-acknowledged personality traits that were causing friction in her new endeavor, and her work to correct them.

It’s a very self-honest piece, a lot like the one done about a year ago on Dominic Orr, Alteon’s CEO (and Ruckus Chairman).  They’re both articles that are pretty brutally honest, and it’s courageous of them to share like that in such a public way.

Both articles mention David Callisch, himself an interesting personality (and a great guy).  He has a brutally self honest quote in a WSJ article a couple of years ago: “All those 0′s“.  David and his team put on the widely successful Alteon Webheads conferences.

Check out the articles, they’re fascinating reads.

The deal was pretty what had been theorized, although I think  the biggest surprise was that Radware has stated (and reiterated in my conversation with Mr Zisapel) that the deal is not a pure customer-list purchase, as many had suspected.  Instead, Radware has comitted to supporting and expanding the existing Alteon line, and not just moving everyone onto Radware boxes.   They’ve also committed to a 5-year product support plan of the current Alteon lines.

The deal isn’t a done-deal yet (the bankrupcty courts of the US and Canada has to approve this) but it seems pretty likely to go through.

Happy New Year everyone!  This little tidbit of gossip was brought to my attention:  Nortel assets are on Radware’s radar.  The article speculates that at the price tag of around $30 to $50 million, Radware isn’t looking at the entire Metro Ethernet asset portfolio (valued at around $1 billion).

I think it’s pretty obvious that it’s just going to be the Alteon ADC group and assets that are going to be purchased.  Alteon hasn’t kept up with the other Enterprise vendors in terms of feature set, but they’ve got a pretty good following.

At a $30 to $50 million price tag, that represents a paltry .5% return (1/200th) on investement of the $8 billion dollar price tag that Nortel paid for Alteon in 2000, during the height of the dot-com-boom.

Ouch.

Nortel hasn’t had much luck lately.  Their market cap is around $150 million (that’s with an M), having left the realm of the billion-dollar businesses a while ago.

Talk amongst yourselves.

The driving force behind Layer 7 persistence (keeping an individual user tied to a specific server in a server group based on HTTP headers instead of IP address) was the dreaded AOL Megaproxy issue.  AOL had the nasty little tendancy of routing all web traffic through a couple of mega proxies located throughout the US and Canada.

This caused a problem with the previous method of persistence, which was to base it on source IP address. Typically, one IP address equaled a single user.  However, with AOL, you could have 20,000 users coming from a single IP address.  The load balancer would think it’s a single user, and if you had 300 servers ready to take orders, all 20,000 users would go to one.  That situation has happened a few times, and it’s hillarious, so long as you aren’t the company with the 300 servers.

I still teach that mega proxy problem, mostly out of muscle memory.  But I stopped to think about it, do we really have a problem with megaproxies anymore?  Does AOL even do this practice, and even if they did, is AOL represent a significant amount of traffic?

The answer to the later question is almost certainly no.  AOL has seen a dramatic drop in subscribers, and most people connect directly to the Internet through their cable modem or DSL provider.  And I don’t know of any major Internet provider that utilizes proxies for their users Internet requests.

Layer 7 persistence is still applicable to situations where you may have multiple users coming from a single IP address (such as a small client base coming from a handful of offices, with each office using on public IP address), but I wonder what doing Layer 4 persistence would do to a major site these days.  I’m thinking, not much.

What do you think?