Synthetic Monitoring
I’m pleased today to introduce you to Palak Patel, the newest member of Apica. Palak joins Apica as the Vice President of U.S. Sales and will be leading the company’s U.S. expansion efforts. Before joining Apica, Palak held various technical and sales-related positions at Akamai. Check out the press release we issued to read more about Apica’s newest team member.
Sara, a member of our blogging team, recently sat down with Palak to gather his thoughts about joining Apica and where he thinks the cloud computing, performance monitoring, and load testing industries are headed.
Read on to see some of Palak’s insights.
Sara: Hi Palak, thanks for being with us today. Why don’t you start by telling us a little bit about your background?
Palak: I spent the last five-plus years at Akamai Technologies, which is the largest content delivery network, and arguably one of the first commercial cloud providers. Akamai was founded in ’98-’99. It really is an overlay to the Internet, with 20 to 30 percent of all Internet traffic running over its network on any given day, and has become, to a large extent, part of the fabric of what makes up the Internet’s infrastructure today.
That’s probably not going to go away, so the content delivery networks are that next layer on top of ISPs. And then, as we move further out, there will be newer technologies and protocols that will layer over this infrastructure — for example, WebSocket as a protocol. And the challenge is how to best monitor and have visibility into all these different layers.
And that’s where a company like Apica comes in, allowing us to have visibility into how our applications are performing from different geographical locations across rapidly changing environments connecting over various protocols. The challenge is finding the most optimal balance between performance, cost, and agility.
Am I actually receiving the benefits of this large, distributed CDN? How do I identify where I can improve performance through my application? Apica’s tools provide this performance visibility on a global scale.
Another piece of Apica’s solution set is load testing as a service. Historically, load testing has only been performed through tools that are licensed and delivered on premise with significant cost implications, largely prohibiting enterprise-scale load testing for many organizations.
Apica is now delivering a cloud model that seamlessly and cost effectively allows application owners to scale to the types of large audiences that we are now regularly seeing online.
They want to be able to simulate hundreds of thousands of concurrent users prior to launching a new product.
I focused on performance and scalability at Akamai, and I will continue to do so at Apica. However, now the focus will narrow down to visibility into these areas and how you can consistently achieve technological balance.
Sara: How do you think your background experience will help serve Apica’s customers?
Palak: Many of the people I was doing business with at Akamai are also the same people that are looking for performance monitoring and load testing solutions. So these are the people responsible for infrastructure – the application owners. And on the business side, we’re seeing that CMOs are responsible for more of the online spend. In five years, CMOs will start spending more on IT than CIOs.
I have some of those business relationships as well because that’s often who is looking at metrics like conversion rates — especially when you get into the commerce business. Understanding the buying process is important from the sales side, but so is understanding customer needs when it comes to performance. I’ve been talking about that theme for five-plus years now.
Sara: Great. What areas of application testing do you think companies are doing well and what aspects do you think organizations need to improve on?
Palak: Customers probably need a lot more education in stress testing applications. That’s because the infrastructure of the web is rapidly evolving from dedicated environments to various hybrid environments where organizations are getting a lot of their computations from machines that they don’t own, which has different implications on their apps. A lot of them are new to this, so they aren’t familiar.
It’s a lot different in the enterprise market than the consumer market and the startup scene. For example, a lot of businesses are now built on cloud providers like EC2. If you think of all the new apps, like Path, Pinterest, and all these new exploding consumer sites, their infrastructure is basically rented from Amazon.
That’s a little different than the enterprise world where they’re trying to catch up. They already have CAPEX and years of experience in building these applications. So load testing, and how to test with the cloud, is definitely an area they need to improve on.
Load testing is more of an art than a science, especially when you’re talking about applications that are plugged into various databases, load balancers, and CDNs.
Our team has a ton of experience in working with these different types of architectures, so we’ve seen where the bottlenecks typically are and can troubleshoot much more quickly. But overall, I think functional testing is where the market is. They know their applications and they can function with that, but when it comes to large-scale load testing, I think that’s where expertise is needed.
Sara: How do you think companies are doing in general when it comes to monitoring their websites? And what could they improve on?
Palak: Right now, mobile monitoring isn’t particularly done well. As an industry, we’re all evolving from synthetic mobile monitoring, but deep mobile network performance testing isn’t fully baked yet. I think we’re getting there, but historically the way mobile and tablets worked was by breaking them into fragments. We were dealing with mobile sites or an “m.” site, which is completely different than your website.
More and more now, though, with smart phones and tablets, users are able to render a full-browser experience. So it’s more about testing how well the website is performing through mobile networks than the true experience in terms of speed. Because mobile sites have been so much lighter-weight historically, the speed has been relatively fast.
It’s very easy to pull up a 25k page or a 50k page on a phone browser, especially an iPhone. But now that you’re getting fully dynamic pages and apps on mobile sites that are completely dynamic and talking to servers, we really need to dig deeper into the mobile networks and have agents within the network. And, frankly, use and monitor different technologies.
I think we’ll start to see a shift in how mobile apps are delivered to these devices, and they’ll start using protocols like SPDY and WebSocket more often. That’s why Apica is starting to adapt and monitor these technologies as well, because that’s where this industry is shifting.
Sara: What do you think the recent switch to IPv6 means to the industry and what does it mean for Apica customers specifically?
Palak: I think Akamai, Google, Yahoo!, and a few others just publicly released support for the IPV6 initiative, and all the major browsers are supporting it. It boils down to the fact that we ran out of IP addresses. And the new version is going to create a whole new slew of addresses and potential new technology protocols to leverage and support.
This will have a huge impact operationally and will uncover inefficiencies, but overall the Internet will improve and continue to thrive as an enabler of the next generation of disruptive technologies.
Sara: From your experience with Apica and Akamai, what do you think are the most common oversights that companies make when they’re moving to the cloud or deploying their mobile technology in the cloud?
Palak: It’s not as easy as just throwing it up in pieces. Provisioning in the cloud can be challenging. You don’t want to rely too much on the cloud when you already have infrastructure. There are some things that are built to rent out and there are obviously some things to keep in-house. So I think the hybrid model is the way to go. One of the mistakes with this is relying entirely on one vendor.
If there’s a huge EC2 outage, your business and your site go down. This continues to happen, as it did recently. That’s obviously a big mistake. So like anything, you want to have hedging strategies when deploying to the cloud. But then again, there are huge advantages — it’s like a virtual sandbox.
If you want to test out certain features and things like that, you can spin up a bunch of servers and be able to do that. So I think it’s a give and take. There are advantages but you still have SLAs and security and compliance requirements, so maybe a dedicated hosting environment is imperative. So that’s something to think about as well. We don’t see the banking industry, for example, adopting cloud as aggressively as consumer companies.
Sara: Do you think that these recent cloud outages are a sign that we’re managing the cloud well? And what steps do you anticipate we’ll need to take in 2012 and 2013 that we have yet to deal with?
Palak: IBM and Verizon are trying to position themselves as the enterprise cloud, so users think there aren’t outages and they won’t have to deal with that. Whereas Amazon has gotten a head start on everyone and opened up its platform. If Amazon’s cloud is down, it’s down — companies take the risk. That’s why so many startups have been able to iterate so quickly. And that’s been the Facebook model — just throw things out there and see if it works.
I think as it evolves, Akamai and bigger players like Verizon, IBM, and even Google, will add stronger SLAs around their computing platforms. Then we can start talking about moving everything virtualized in these environments. But until then, it’s still that sandbox type of environment.
Sara: What drew you to Apica in the first place? And what do you think makes Apica’s approach to cloud testing and testing mobile apps unique to other solutions?
Palak: I think the ultimate vision around Apica is what’s most exciting. Monitoring and load testing as platforms are already established solutions. There are many players out there, not as many on the load testing as a service side, but certainly with monitoring.
There are just plenty of solutions, and we’re trying to keep up on the number of technologies we’re supporting. And, as the web evolves, we keep up with the newer technologies and are able to monitor and test against those.
But the unique thing that Apica offers is that we’re opening up our platform for you to build plugins and applications to get more insight. That includes lower-level data, since we’re not burdened with storing all the information. You can pull anything you want and store it yourself and then build applications on top of it.
It is like what Amazon has done with EC2 and the APIs, but instead providing a full data set of all usage so customers can query and curate as they please. We are trying to open up the platform so you can perform computing on your own to create anything that ties into our performance data.
For example, Google announced a new feature for Google Analytics. It allows you to conduct browser side analysis so you can see what percentages of your viewers are browsing what parts of the page. Imagine if we tied that into Apica.
So, for example, I go to a commerce customer and say, “Well, from New York and from LA, this is how customers are viewing your pages, and this is how performance impacts their behavior from different geographies.” That type of insight will influence product placement, etc. That’s incredibly powerful.
So if you look at the road map for Apica, I think we have a lot of compelling applications in mind. But the great thing is there are many more that I don’t know of that customers and developers will come up with. That’s the beauty of having an open platform.
Sara: Thanks again for your time Palak, it was great speaking with you.
Palak: Likewise. Thank you.