AppDynamics debuts on-the-fly fix functionality for enterprise apps

AppDynamics debuts on-the-fly fix functionality for enterprise apps

Summary: The latest release from the application performance management, or APM, company automates repairs in the application layer -- with a little help from real-time analytics.

SHARE:
appdynamics-logo-340px

San Francisco-based AppDynamics, which develops Application Performance Management (or APM) products, announced this morning a new capability that allows it to fix, not just find, problems in the application layer.

IT Operations professionals are used to automation in terms of server tasks -- deploying operating systems, applying patches -- but they tend to avoid automating activity in the application layer because it can cause headaches. 

The company's Application Run Book Automation product intends address this by identifying, then resolving, issues on the fly with the help of real-time analytics.

The idea: fix simple things faster.

The functionality is already in use by Edmunds.com, Expedia, and Orbitz, all existing AppDynamics customers.

The company calls this capability "disruptive." To tease that out, I rung up founder and CEO Jyoti Bansal.

Here's what he had to say.

ZD: Give ZDNet's readers a quick introduction to your company.

JB: We're a five-year-old company started in 2008. We help people manage large web applications -- if you're Netflix or Priceline or Expedia or Edmunds, that's important. We have more than 500 customers. There's also Orbitz, Staples, Intuit. Companies where a significant part of their business is running [in the cloud].

There's about $2 billion a year in IT spending on this, according to Gartner.

There is a major shift in how applications are being built. For the last 10 or 12 years, we've had very monolithic applications. Now there are highly distributed architectures.

Take Netflix -- there is three, maybe $3.5 billion in revenue that comes through that application. Any downtime, instability, and so forth can become difficult to manage.

The legacy software vendors are incumbents -- IBM, HP -- who built their products before cloud and virtualization happened. Our mission is to build the next-generation management platform.

ZD: Where's the action in this space?

JB: Software is leading the world. Any business, it doesn't matter what you do -- you have to become a software applications business. If you're a bank, insurance, telco, bookstore. It's a big, big trend [in terms of] how much of the business is moving to online applications.

Mobile is an equally as large disruptor. It's opened so many new application opportunities. We do see mobile mostly as additive. For some businesses, mobile is replacing -- like what the guys at Square are doing at the point of sale. But for most businesses, it's additive. There are use cases that were not possible before mobile.

Things are much more distributed. The same application stack, set of servers have to be able to serve the physical retail bank location and the websites. They have to be much more agile. The scale is different. The user interface in mobile is changing rapidly. They used to change it once a year; now it's faster. And the scale is much larger. Scalability, agility, distributedness.

The approach we take is, the best way to manage these applications is the ability to instrument every line of the code. We can watch it in real time to watch what's going on. Given what's happening in terms of how people are building apps, it's a very exciting space right now.

ZD: Surely there are growing pains.

JB: Of the challenges, the number one is lack of visibility. Unless a customer calls and complains that things are slow, or the user experience is bad, a lot of these businesses have no idea what's going on. They put applications in production, and it becomes a black box. A billion-dollar business that's a black box. It's surprising. The first generation set of tools and management services aren't working in the new apps.

The second thing is that things do go wrong. They'll never be 100 percent perfect. You don't want to wait six to nine months to release new code to deploy a new feature. You want to do it every few weeks. As you become more agile, things go wrong. It's not about things going wrong, it's about what you do when things go wrong. Facebook, Google -- they are able to manage applications at a humongous scale because they have perfected this part: automating figuring out what's going wrong and fixing it.

We can instrument everything from a user's click. One of the new areas we're announcing as part of this product launch is taking it to the next level -- providing insight into what's going on and find the root cause of the problem and automatically fix it. If something is slowing down, our technology automatically fixes it to reduce the time users are impacted -- in a few minutes, or even a few seconds in many cases. We can automatically roll something back. Capacity problem? We can automatically scale the application.

ZD: Is this really a technology problem? Often, human execution and prioritization is as much to blame for irritating end user issues.

JB: It will be a gradual journey. "I have no idea what's going on. When something happens, I need to put an army of engineers on it to fix it." That's where people are. You need to get visibility. You need to know what's wrong. And in many cases, you can fix something before it becomes a problem.

The second step will be that you don't even have to know -- the software will fix it. Frankly speaking, this won't be solved overnight for all problems. It may be solved for a quarter of the problems. There are certain classes of problems that cannot be automatically be fixed, and they still need visibility into what's going on to fix it.

The scale of these applications is so big -- people used to have 10, 20, 30 servers to serve an application. Now they have 5,000, 10,000, 20,000 servers. Humans can't handle that. There needs to be some automation.

ZD: This need for visibility and granular data: is it being driven by economic reasons, or is technology pulling enterprises into the future?

JB: People are struggling with visibility, but they do struggle with that. There's too much business running in these apps to accept that it's just a black box. A lot of companies are trying to piece together this information over a year or two. If there are the right management solutions in the market, people will [change their ways].

When cars were first built, people didn't think of the instrumentation on the dashboard. There was no visibility. Once you start using it every day, you need to know what's going on. If something goes wrong, you want a light to know what's going on. You start with getting something out, and as you use it every day, you realize the importance of visibility and insight. People are going through that transition today.

The flip side is, it scares people to go through this transition. Traditional businesses talk about why they aren't adopting the public cloud yet, they talk about a lack of good management technologies. They won't take the chance of putting a large application on new technology until they know they have the visibility.

ZD: You have several big name, web-first clients. Now that they're household names, are they as aggressive as they once were about trying new technologies?

JB: It depends business to business. If you look at Netflix, they're moving from being a DVD shipping business to a movie-streaming business. They essentially have to become a startup again -- fast, nimble, agile. They're focused on moving fast. It's difficult to do.

Many businesses don't go through transition all the time. They stabilize. Customers aren't expecting major disruptive features. People aren't expecting TurboTax to change -- they just want it to be stable. For companies to be competitive, they have to be very consumer focused. If you have a bad experience, you can leave for a competitor very quickly.

ZD: And I must ask: are you running your own business on your own products?

JB: We instrument all our applications through our software. We use it as the primary management for them. We do it for two reasons: one, we test it -- we know we'll find bugs before customers do. Second, we do believe that there are not too many other good solutions out there.

We need to provide a really good user experience. We are instrumenting very large, high-volume applications out there. We are capturing a lot of data that's coming to our services. We are instrumenting and watching 50 billion connections a day. Our own application is very high volume, distributed, agile. It helps our operations team and developers to see how our end users will be using our software. Once they see that, smart engineers figure out very quickly how to make software better.

We went from 80 employees last year to more than 200 this year. We just raised $50 million in financing a month ago. (In a Series D round that included Greylock Partners, Kleiner Perkins Caufield & Byers and Lightspeed Venture Partners. --Ed.) There's more expansion to come. As that happens, the scalability of our application will be a challenge.

Topics: Enterprise Software, Data Management, Start-Ups

Andrew Nusca

About Andrew Nusca

Andrew Nusca is a former writer-editor for ZDNet and contributor to CNET. During his tenure, he was the editor of SmartPlanet, ZDNet's sister site about innovation.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

1 comment
Log in or register to join the discussion
  • the AppDynamics copy book I believe

    The Complexity and Challenges of IT Management within the Cloud
    http://www.jinspired.com/site/the-complexity-and-challenges-of-it-management-within-the-cloud

    "Bigger and faster growth & change rates sounds like a disaster. How can we reason about the problem space when we are being forced to see light years into the future (in terms of the computing speeds) and in a space that is vast (in terms of our mental capabilities) whilst expanding and contracting on demand for causes we still have very little insight and control over? How can we even reason about the present when the past seems so distant and different in this regard. Remember BIG data is not a model that is something we have to mine from it.

    The answer lies in emergent computing and adaptive control that is local and immediate. Local in that observation, judgement and reaction are collocated with the normal processing via embedded controllers and sensors weaved into applications (at runtime). Immediate in that the time interval between measuring, sensing and signaling (possibly to a remote station) and the actuating is at the same resolution of the underlying task/transaction processing that is being monitored, managed and controlled."

    And the solution has already arrived for many issues at scale

    http://www.jinspired.com/site/jxinsight-opencore-6-4

    "Whilst many companies today are increasing the time and expense spent in collecting, storing and analyzing application performance monitoring data the results viewed in terms of improved availability and reliability are invariably disappointing because monitoring is not management. To actually manage applications, operations need to go beyond passively watching an application crash as it uncontrollably (and with little regard for its safety) reaches and exceeds internal and external resource limits and barriers. Today sadly the most commonly practiced operation performed by management staff tasked with resolving a troubled or failing application instance is to kill and restart the runtime.

    With JXInsight/OpenCore 6.4 we are offering three powerful metering extensions based on our adaptive control valve technology that dynamically optimize the performance of an application in terms of throughput and response as well as safe guarding its execution from exceeding limits it is not fully aware of but that can be sensed via past, present and predicted activity metering."
    william.louth@...