Boundary app monitor becomes a fortune teller
One Dashboard to rule them all
How do you solve a problem like poor performance in distributed IT stacks? If you're application performance manager Boundary, you do it by teaching your tech to not only analyse traffic on the network layer, but also listen to app-specific alerts from typical management systems, and use this to predict where bottlenecks are going to occur.
To do this, the company has dramatically expanded the range of software tools its app monitoring platform can listen to when analyzing performance, and has introduced new predictive features for its analytics package, Boundary announced on Tuesday. The company's as-a-service tech now includes inputs from AppDynamics, Nagios, New Relic, Opscode, Plexxi, Puppet Labs, and Splunk – along with gaining a redesigned management console.
These upgrades pair Boundary's network analysis tech with inputs from the above products to give admins a way of looking not only at the layout and behavior of their network, but also gain deeper insight into the applications on it.
It can also now offer early-warning alerting, which will let it telegraph potential developing problem areas to admins before problems began to slow down data flow.
If a datacenter were a city, then the inputs mean Boundary has the ability to monitor not only the roads, subways, and sidewalks (the network), but also what's going on inside the buildings (the applications).
"We can see problems others simply cannot see," Boundary chief Gary Read told The Register.
The company's technology works by monitoring the flow of information between applications and infrastructure via software "meters" deployed on server instances. The as-a-service technology can work with both cloud and on-premise environments.
"The dataset we are pulling from is a network flow dataset," Read says. "[Boundary] was not ever really designed to be monitoring and managing the network, but instead to use the network to monitor and manage the apps traversing the network."
Unlike most other application performance management services, Boundary maps and mines the network of data flows between applications and gear to give a view of performance, whereas other software packages are either app-specific, making the technology less granular, or operate on a significant delay of minutes to hours for information updates, versus Boundary's single-digit-second latency.
It already works works closely with Amazon Web Services environments to give admins a rough view of how the AWS network is functioning – no surprise, given that Boundary's cofounder Benjamin Black came up with the seeds of what later became AWS EC2. Along with this, it supports other clouds such as Windows Azure, Engine Yard, the Rackspace Cloud, SoftLayer, and Google Compute Engine. It can ingest any data outputted in an event or alert format via a variety of mediums, including RSS.
The company launched in April 2012. As of August 2012 it had around 50 paying customers, and decided to launch a freemium version. In the intervening year, things have apparently gone well for the company, with it now claiming 95 paying customers, over 1,000 freemium punters, and touting the fact that its platform is chowing down on 7TB of data daily. Read thinks the company will "probably" have 250 to 300 customers in a year's time.
Pricing for the service is free up to 1GB of analysed data per day, then goes up to $495 per month for 5GB per day, $995 for 15GB, and so on. A bit of napkin math means that even if each of the company's freemium customers are consuming their max, that still leaves 6TB under paid management, which means the company could be turning over somewhere around $400,000 per month, assuming price reductions and a few super-sized (and heavily discounted) accounts.
Boundary's proposition pairs its so-far-unique network topology analysis with a grab-bag of handy features that punters can find in other products. From what we can tell, reconstructing Boundary's capabilities outside it would require a bunch of different software tools, including cloud monitoring from companies such as Cedexis for the early-warning component, Compuware for cloud monitoring, or New Relic for app monitoring.
"You cannot do these things as well, as fast, or with as broad a coverage without having per-second network data availability to you," Read says. "That per-second network data is the engine and the other pieces are the car being built around that engine. Could you build a lesser car with the different engine? Yes, but it wouldn't be as fast and it wouldn't be as good."
Instead, Boundary thinks it best if its network tech is paired with app-specific management platforms. Read sees the company's technology as an additive tool for an admin, rather than something to replace others. ®