Yet, that’s inevitable for companies such as Google, Microsoft, Amazon, Yahoo, Facebook and others. Even as Netflix outsources most of its IT operations to Amazon, Amazon must figure out how to economically scale its business — and having the old industry standard of one systems administrator managing 500 or maybe 1,000 servers isn’t going to let AWS keep dropping prices. This is why a partnership between data center networking hardware company Plexxi and network monitoring company Boundary intrigues me.

The two companies have created a means for customers who have Plexxi gear installed in their networks to use data provided by Boundary’s monitoring service to automatically adapt the network in real time to the demands of an application. So if Boundary determines that the network flows it’s monitoring are slowing down because the database isn’t feeding information fast enough to the CPU, Plexxi can widen the available bandwidth between those two units until the bottleneck is resolved.

Having covered IT for over a decade, I can tell you that I’ve seen a lot of marketing around this stuff, but once you dug deeper, the caveats and clunky integrations stole a lot of the possible benefits. But the ability to share information using APIs, plus both companies hewing to the idea that they are creating open platforms has made integration relatively simple.

Boundary’s service discovers a latency problem.

Plexxi begins to address the problem by changing the physical network.

“We’re two companies that share a vision around the abstraction of what we call affinities — stuff that wants to be grouped together,” said David Husak, Plexxi’s CEO. “Boundary can look at the activites of those services and derive affinities, and then Plexxi can do something with it. This is not exclusive between Boundary and Plexxi, but we’re both performing and rallying around this idea of affinity abstraction and that’s what makes it powerful.”

That doesn’t mean there aren’t caveats. This only works for those using Plexxi gear, so people running in cloud environments such as Amazon’s EC2 can’t do this yet. And Plexxi’s gear is pretty new on the ground, so it’s unclear how big the customer base it. One also has to subscribe to Boundary’s service, but there are hundreds who use the free version and 80-something customers on the paid version today.

And if this doesn’t work for most people, that might be alright because startup Lyatiss, a company I covered last month is trying to do something similar with its software. It’s stuff will work in Amazon’s EC2, although so far it only has the monitoring component as opposed to the automatic scaling element. However, thanks to APIs, an acceptance that vertical integration doesn’t work in servers, networking or in storage, and the demands of scale out data centers we may be closer than ever to application-aware infrastructure. The kind of IT that when it’s broken (or approaching broken) can right itself.

That’s pretty cool, and it’s exactly what we’ll need if we want our computing to scale.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Cloud computing infrastructure: 2012 and beyond
• A near-term outlook for big data
• Q4 Wrap-up: SOPA and the future of digital content

Whereas application-performance-management providers such as AppDynamics and New Relic help clients zero in on problematic code inside applications, Boundary looks out for issues slowing down the network, including the application itself.

Based in San Francisco, Boundary has more than doubled its workforce, going from 12 employees to 28 since Read joined the company last January. Since launching in November 2011, Boundary has grown its clientele to 600 customers, 76 of whom pay for the company’s services. And the amount of data flowing into the company’s infrastructure from clients has grown substantially, from less than half a terabyte to more than 5 terabytes daily. For comparison Facebook stores 1 percent of that amount every 24 hours, according to a November post from the social networking company.

Amount of network and application data Boundary has processed per day in recent months

Annual revenue has gone up, too, although Read declined to provide figures.

Some infrastructure hiccups have accompanied all that growth.

“So definitely as you continue to scale the system, it’s very difficult to test a system like this to unlimited scalability, and so as you continue to push more and more data then it will show itself in different parts of our platform, and different tiers in the application may start to run out of horsepower, or we may start to hit certain limitations in particular areas,” Read said. “We’ve seen that twice already … . In one case, we had to use solid-state disks instead of physical disks. … In another case, we’ve had to add more servers and more processing thru that infrastructure, to deal with us starting to get to capacity limits in what we could be processing.”

More challenges could lie ahead, Read said.

As more clients sign up for monthly or yearly contracts, more data is entering the equation. Boundary could open a second data center, exclusively for paid users, to offer better service, Read said. A service that processes one gigabyte is available to users free of charge.

“With just the sheer volume of data being dealt with so very, very quickly, and so much insight given quickly into that data, this is something where we’re really starting to move into ground that’s never been attacked before,” Read said. And as that data grows it seems so far Read has been able to scale Boundary’s infrastructure. So far he said that the amount the company spends on infrastructure is declining as a percentage of revenue, which is to be expected since the company over-provisioned on its hardware in the early days.

However, as time passes and Read contemplates adding more capacity, he’s confident that Boundary can continue to scale both the infrastructure and its business model. Given how useful network monitoring can be for cloud-based applications Boundary should prepare for more customers and more competition, such as the newly launched Lyatiss.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Cloud and data third-quarter 2012
• A near-term outlook for big data
• The fourth quarter of 2012 in cloud

Okta is part of the cohort of companies that have seen the shift to the cloud as an opportunity to rethink the way legacy enterprise software is built. Other players include Boundary, Aryaka and Workday. Each of these firms has seen the way that cloud computing and software as a service can disrupt the enterprise IT market and have made bold bets building out new services that take advantage of the decentralized nature of the cloud, cheaper compute or new business models. The real question as the startups gain customers and credibility is which of the IT giants will buy them and for how much of a premium?

Okta, for example, offers a centralized identity management platform that competes with services such as Microsoft’s Active Directory and Oracle’s Identity Management suite. Okta’s software works to establish user identity on both cloud applications but also for legacy and on-premise programs. This saves Okta’s customers from having to use two different identity management products for employees. Todd McKinnon, CEO, Okta told me in an interview ahead of the announcement that it has seen a lot of success recently stealing customers from the established vendors, and that the funding will help it continue to invest in connecting its platform to older, legacy applications that its customers want integrated into the Okta software.

Otka says it has added more than 140 enterprise customers and 300,000 users to its service in 2012, including Allergan, BMC Software, Clorox, Groupon and National Geographic. Okta has developed relationships with more than 2,000 cloud applications, all of which come pre-integrated in the Okta Application Network, and counts leading enterprise SaaS companies such as Jive, Lithium and SuccessFactors as customers.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Social 2013: The enterprise strikes back
• Cloud and data third-quarter 2012
• A near-term outlook for big data

The idea is that Boundary has analyzed the performance and habits of a customer’s network so when that pattern changes, Boundary can send an alert. The company is ingesting 350 Mbps of inbound networking traffic on its 550 customers which it then processes in real time. Boundary has its own custom-built, real-time data processing engine, that takes this influx of second-by-second data and combines it into dozens or hundreds of terabits per second of traffic for analysis. A Boundary spokesman notes that this current rate represents a fraction of the startup’s total processing capacity.

Indeed, both the product launch and the data flows are topics I’ve discussed with Boundary CEO Gary Read (pictured) in the past. We’ve outlined the merits of real-time network monitoring that can track multiple data points every second as well as discussed what that will mean in terms of the amounts of data that Boundary will have to process. With 550 customers, Boundary is analyzing 600,000 events every second and the company is adding about 10 to 12 paid customers each month.

To analyze this data, servers monitored by Boundary report data to its SaaS platform, sending up information about the network and application flows they see passing through. So the 600,000 events per second refer to the network flows that Boundary observes each second for our customers inside their networks, and as they communicate with their customers. As Boundary grows to observe more flows between data centers to end users, and through mobile networks, it gains a broader picture of the health of cloud providers and the Internet as a whole. It’s somewhat similar to what startup DeepField Networks is also proposing, and is an essential step in creating accurate monitoring for federated applications built in cloud environments.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• AWS Storage Gateway jolts cloud-storage ecosystem
• Dissecting the data: 5 issues for our digital future
• Is Tiered Data Access Unfair, or the American Way?

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Takeaways from the second quarter in cloud and data
• Cloud computing infrastructure: 2012 and beyond
• Amazon’s DynamoDB: rattling the cloud market

Boundary CEO Gary Read

Gary Read, the CEO of Boundary, said the company aims to use its ability to see second-by-second networking stats as a way to understand and indicate the health of an overall application. It’s analogous to Splunk using machine log data to tell the same thing. And given how much network health can impact overall application health, especially since most people access their software as a service or build apps that are reliant on a variety of API calls, using network information to determine performance can be an early warning system for application failure.

That’s one of the things Read and I spoke about a few months ago, when I explained what Boundary and real-time network monitoring has offered customers so far. For the most part it helps customers see more data and get it faster because it checks the network and offers updates every second, but that quantity and access to fast information can also help customers save money if they are building on platforms such as Amazon Web Services or Google’s Compute Engine.

Users of the service say they can quickly see if they’ve built their applications in a manner that will cost them a lot just by tracking the network flows. The key here is that Boundary is assessing information every second, in a way that’s perfect for a cloud-based world where the demand for services can fluctuate on a second-by-second basis. Three months after launching its application monitoring service, Boundary is tracking 20 billion records per day. That’s a lot of data, and Read only hopes those records will continue to grow.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• The evolution of consumer-media cloud storage
• Infrastructure Q1: Cloud and big data woo enterprises
• Controversy, courtrooms and the cloud in Q1

Boundary CEO Gary Read

What if you could know everything about your network? What if instead of getting snapshots — albeit very rapid snapshots — you could see the path of every packet and run basic analytics on that stream of data in real time? It’s the difference between watching a Pixar cartoon as opposed to viewing a flip book. And that changes things.

That’s why I was so excited last year to learn about Boundary, a startup that has raised $4.1 million and now has 21 paying customers after about 6 weeks of making its subscription networking monitoring service generally available. The company lets customers see their network operating in real time — every packet and every flow. Each day it gets about 200,000 inbound records per second and generates about a terabyte of data that is processed through its proprietary data store — built using a combination of Scala and Erlang.

The startup is cool, but in a chat with Gary Read, the CEO of Boundary, we discussed what customers have done with the platform and how real-time monitoring and embracing all the data instead of some of it has allowed customers to see more, see faster and save money.

Seeing more at DNSimple. Anthony Eden, the CEO of DNSimple, is using Boundary’s service to monitor traffic flowing into his DNS provider. He had tried other services such as those from New Relic, but they didn’t give him the detail he needed to understand the traffic hitting his servers.

The new visibility let him spot interesting traffic patterns, especially related to requests coming from China. Eden explained that he didn’t know if the traffic was malicious; it just looked different. “Crafted” is the word he used. The traffic patterns were subtly adjusted — something other DNS providers had seen as well. So now Eden is faced with a new traffic pattern and plans to keep an eye on it. Luckily, by watching everything in real time, problems can be detected earlier.

Seeing faster. Another customer (who preferred to be unnamed) says the service helps them detect network problems and attacks about five minutes ahead of what other software allows, thanks to the all-encompassing view of the packets. And with the speed with which a network problem can go from anomaly to all-out failure, five minutes could be the difference between a site slowdown or something like Amazon’s massive outage from 2011.

Saving money. Having access to more data can also help developers quickly update their apps and save money. Another unnamed customer who had moved applications from Google Apps to Amazon Web Services realized immediately after the move that their app wasn’t performing well. After hooking it into Boundary they realized the way the app was accessing an external DNS provider was costing the app developer more money. So the developer switched to Amazon’s in-house DNS lookup service and rebuilt the application to optimize it for Amazon. In the process they estimate they have saved $15,000 per month and seen performance speed up within a few hours.

To be clear, other networking monitoring applications would likely have helped in each of these cases, but the key for customers appears to be the speed and the amount of information Boundary can parse. Of course, not everyone thinks it’s necessary to look at everything, and for some apps it may never be worth it. They won’t ever need a Pixar-style animation.

Plus, there are plenty of questions about how well Boundary can scale. If 21 customers generate 1 terabyte of data each day, imagine what happens when it has 100 or 1,000 customers? For now, Boundary starts culling data at the one-day and one-year mark, so at 1 month you might have minute-by-minute data and after a year you only have hourly data. But Read says customers could pay more for more storage. It’s early days for the company, but given the movement to a real-time Web and the speed at which things can change online, Boundary appears to be a necessary service for those who want to keep up.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• How devops can reduce cycle times
• Cloud and data third-quarter 2012
• A near-term outlook for big data

Benjamin Black, Boundary co-founder and CEO

Boundary, a startup that aims to rethink network monitoring for the cloud, has raised $4.1M and officially launched. The San Francisco-based startup, founded by Amazon Web Services veteran, Benjamin Black, is launching today with the news of its funding (that was raised in January) and promises that the service will be generally available during the first quarter of 2012.

The 15-person company, which was formed in late 2010, raised the funding from LightSpeed Venture Partners and private investors associated with Amazon, said Black. Boundary is rethinking network monitoring for distributed and federated cloud-based applications under the theory that in the cloud, the network becomes your IO and not knowing what’s happening there could kill your business.

“If you are not monitoring your network and the packets and bytes moving back and forth then you are not running your business effectively and missing out on numerous opportunities to operate things more efficiently and troubleshoot outages and identify outages,” Black says. “And all of that depends on high-resolution, real-time network monitoring.”

The service, which is hosted on Boundary’s own infrastructure (Black said he didn’t want to deal with the issue of figuring out new code and troubleshooting it for the AWS infrastructure at the same time) deploys agents on the customer instances that report back to the Boundary software.

The software consists of a real-time streaming engine that analyzes the network traffic every second and transmits the information to a dashboard. It’s possible to monitor the usual information about speed, ports and the route of a packet, but one can also measure other data such as the geographic locations of packets, which might be useful for detecting certain security threats.

It also has an API so folks can build applications that tap into the service that could be tailored for those who are not network engineers. For example, building a dashboard for a CFO that shows how much money is lost because of network lag times might help get the executive team behind a vendor or equipment switch. Another use case, might involve setting up an audit for ensuring traffic complies with national laws or industry certification standards.

Folks can install the service manually or have it load as part of a Puppet or Chef configuration. Each agent has its own credentials so it’s not bound to a VM. That means if an Amazon instance disappears for example, the agent can get certified as it pops up again on a fresh virtual machine. Most old-school network monitoring software requires the network to stay fairly static, something that doesn’t happen in a virtualized world.

Black notes that today the service only monitors the network and can’t take action, which would be the logical next step for a tool such as this. For now customers, and Boundary has a few already testing the service, will pay for each monitored server with the exact price being worked out during the beta period occurring between now and January.

In general Boundary is riding one of the next big shifts in IT as a result of virtualization — how the network will adapt. As the network becomes more dynamic and flexible, it will require new ways of monitoring it and controlling it. Boundary believes it can help solve the first problem with its software.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Cloud and data third-quarter 2012
• Quality of the cloud: best practices for ISVs
• Migrating media applications to the private cloud: best practices for businesses