When my previous company, Wily Technology, was about 6 years old, our application performance management software, Introscope, was starting to get widely adopted in some large companies – insurance companies, banks, telcos, you name it. Typically, the customer would begin with success in a single application environment. With intent to replicate success, they would then purchase a large number of licenses in bulk through an enterprise software license agreement (usually well north of $1 Million) with the intent of deploying Introscope everywhere. This is typical of the enterprise software business model.
And here’s where the business model hurts the customer. As I mentioned in my previous post, most enterprise software is poorly designed, especially as it relates to deployment, management and usability in large environments. One of Wily’s larger customers (a big bank you’ve heard of) complained that it took 2 of their IT admins 5 days to configure Introscope to manage one web application. That’s an enormous cost, and it doesn’t scale in an organization with more than a thousand applications that they want to manage. They issued us a challenge that they called “2-3-100”:
They wanted 2 sys admins…
…in 3 days…
… to fully deploy instrumentation and management for 100 web applications.
Unfortunately, the fundamental architecture of Introscope (and virtually every other APM product in the market today) made this impossible to achieve – too many configuration options, too much customization needed, too many cooks in the design kitchen, too much rope for the customer to hang himself with. So the solution Wily and its industry peers came up with is a favorite band-aid in the Enterprise Software world: Professional Services. Not only would our customers have to shell out huge bucks for the software, they would also have to pay up to $400 per hour to get one of our smart technical guys on-site to configure the software to make it do what the customer wanted.
Since I went after my alma mater Wily in my previous posting, let’s spread the love and compare the setup process of New Relic versus IBM’s Tivoli Composite Application Manager, which (after you read all the gobbledygook) is really another production application performance management system that is an erstwhile Wily imitator.
Full Production Setup with New Relic:
- Sign up for a free account. Your email address is all we need.
- Add our agent to your app server’s class path. We’ll auto generate the config file with the right settings, specifically for your account.
- Push your app to production, running exactly as it has before. No new servers to deploy, hardware to configure, etc.
- Log in to New Relic RPM. You are done. Zero additional configuration required to monitor, tune and diagnose apps.
- To deploy on another app, repeat steps 2-3.
Full Production Setup with IBM Tivoli IT Cam
Actually I misspoke. It’s actually a pre-production setup, but it still takes 10 days. Don’t even bother trying to do it yourself. Why would you when you can get IBM consultants onsite for hundreds of dollars per hour (plus travel and expenses, natch). Here’s a glimpse of the “Offering”. I love that they call it “QuickStart.” It implies that (a) 10 days is quick, and (b) if you want to do the “real thing”, a larger time and capital investment is required. One of the key challenges the offering addresses is “minimizing the risks associated with ITCAM for J2EE deployment project”. Minimizing risks associated with deploying management software? I thought the software was supposed to reduce risks, not introduce them!
Surely application managers deserve a better experience than this. Who has time and money to waste on getting software installed and configured?
Look, I know its beginning to sound like we really must have messed up at Wily. To the contrary, Wily is the best of the bunch. No, its not that. That’s like thinking that you must have been a real bonehead to have bought a satellite phone for $3000 back in the 80’s. It was a huge breakthrough, cutting edge, and the best option available to communicate from anywhere in the world. But before too long it became obsolete. A better, cheaper, more reliable option came along, called the cell phone. We are creating that same kind of dislocating event in the world of app management.
I would love to hear your comments.