A10 Thunder TPS Review

Enables us to mitigate almost all attacks we detect, and with reduced manual intervention


What is our primary use case?

As a cloud infrastructure and hosting company, we provide public and private cloud services. We use the A10 technology to protect our customers against DDoS attacks against their hosting instances. The latter could be dedicated servers or virtual servers or data storage platforms.

We have technology in place to detect attacks at the border routers of our networks in all 20 data centers worldwide where we operate. If we detect an attack, we reroute the traffic of the IP addresses that are being attacked to the A10 Thunder systems. They drop the malicious attack traffic and they pass through the legitimate traffic to the servers or the virtual servers that our customers use.

We deploy different types, different sizes of equipment. We use these solutions on-premises.

How has it helped my organization?

When we started deploying this and we measured the impact on the number of customer complaints, we saw a significant reduction in the overall number of customer tickets. If customers have an issue with one of our servers, they open a ticket; that could be any outage or a DDoS attack. We saw an overall reduction of 11 percent in support tickets.

But we also saw that we were typically able to mitigate over 98 percent of all the attacks that we detect. That has a two-fold benefit. First of all, customers are happier because their service stays alive even in a situation where they are being attacked. And for us, it has a positive impact on our support team because it has 11 percent fewer tickets it needs to handle. That's especially true since "attack tickets" are not nice tickets to have to handle. It also helped us a little bit in the engagement and the motivation of our support team.

A10 has also definitely reduced the amount of manual intervention required during an attack. Before we had these systems in place, if an IP address or server was attacked above a certain level, we would manually no-route or "black-hole" the traffic, and basically remove that IP address from the internet. That was all manual work, while customers were complaining, and their customers were complaining. People were opening tickets. With this solution in place, all that manual work no longer has to happen. After detection of an attack, the scrubbing is initiated automatically. In the case of a huge attack, we will still null-route the traffic which is going to the IP address under attack, but that process is fully automated. So deploying these systems has reduced a lot of the manual work.

Using this solution we have also, to some extent, detected more small attacks, attacks that we had been missing previously. Before we deployed A10, we did not have any technology in place to detect an attack. Only if a customer opened a ticket did we know there was an attack. But when we started deploying the detection technology and the A10 scrubbing technology, we suddenly saw that we actually have a lot of smaller attacks as well, which were invisible to us previously. That means, most likely, that there were a lot of unhappy customers - or unhappy end-users of our customers' systems - that we were never aware of. That was suddenly fixed by deploying these systems. In all of 2018, we identified about 400 attacks each day, anywhere in our 20 data centers around the world. Many of these attacks were invisible to us before 2016 when we did not have this solution in place.

When it comes to the solution's performance given its form factors, for us, any equipment that takes up space and power is using scarce resources in a data center. The fact that these boxes do have a small form-factor, as only 1RU or 2RU devices, and that the power consumption is relatively low, is very beneficial for us.

We don't deliberately use the solution's machine-learning powered Zero-day Automated Protection (ZAP) but the systems require very little effort to keep them alive and manage them. The automation and the updates that A10 built in result in there being very little work for us to do to keep these systems up to date and efficient in the way they scrub attack traffic. So it's not functionality that we deliberately use, but it's a benefit of these systems, which helps us maintain a low cost of operations and an effective system.

The solution's automation also has the effect that the systems are very low-maintenance. That means that we can free up our people to do other work.

What is most valuable?

The primary benefit that we see from their systems is that their filtering technology has the ability to detect and drop the malicious traffic from the legitimate traffic with a high success rate. That, in combination with the very small effort needed to manage their systems, are the two most important benefits to us. On the one hand, it's the quality of scrubbing, and on the other hand, it's the low total cost of operations for us to keep these systems alive and working efficiently.

The quality of the scrubbing is, of course, what the system is supposed to do. It's the key functionality of the system. That's what we bought the equipment for. And the small effort to manage the systems and keep them alive, of course, immediately translates into a benefit that we have a low cost of managing those systems. That means we can allocate the time of our network engineers to other activities.

If you look at the total response time that we see in our solution, which means the time between the start of an attack and the time that the scrubbing really starts, we typically see two to three minutes. But the majority of that time is actually used by our detection technology, not so much the A10 network scrubbing technology. And then it takes a bit of time to reroute the traffic to the A10 equipment. Once it has been rerouted, the scrubbing starts very fast, so the start of the scrubbing is only a small part of the two to three minutes. In general, we're very happy with the response times and the scrubbing quality of the A10 equipment.

What needs improvement?

For how long have I used the solution?

We started deploying A10 TPS at the beginning of 2016.

What do I think about the stability of the solution?

I've heard no complaints, so my perception is that the systems run very stable.

What do I think about the scalability of the solution?

The solution enables us to scale defenses. We use different types or sizes of equipment. Typically, we start in some smaller locations with the smaller equipment type. When we see that location growing, we typically replace that device with a larger one and we move the smaller device to a new location where it's needed. We move the technology around quite a bit, which is our way of scaling up. The fact that there are different sizes of equipment, all with the same technology and the same processes for managing them, is very helpful for us. If you look at our smallest data center worldwide, it's a location which generates around ten gigabits per second of outbound traffic. That typically means about one or two gigabits of inbound traffic. Our larger data centers generate around 1.7 terabits per second of traffic. That's a lot more. And with one family of products, we can still protect both the smaller data centers as well as our larger locations.

So far, the systems do what we expect them to do and they scale as we expect them to scale.

How are customer service and technical support?

Overall, our experience with technical support has been positive. We've had very few requirements for technical support. I know there's a 24/7 SOC team available to help us with large incidents or attacks which we can't resolve ourselves. But so far, we've never had a need to use that team. It's easy for us because the A10 team lives just a couple of blocks away from us. That makes it a bit easier to communicate.

If you previously used a different solution, which one did you use and why did you switch?

How was the initial setup?

The initial setup was pretty straightforward but we also had very good support from the local A10 team here in the Netherlands. Our headquarters are based in Amsterdam. The A10 Dutch office is just a couple of streets away from us, which also made it easier to work on this together. But having said that, the systems themselves are pretty easy to deploy.

Our initial deployment, back in 2016, happened in what were our six main data centers at the time. The easiest one for us was here in Amsterdam because it's almost next door to our office. The deployment itself, the physical installation and activation of the system, is not really the critical activity. Most of our time was spent integrating the systems with our own administration systems, so that we could deploy automatically. And there was the whole setting of profiles for IP addresses to understand how the detection should work and how the scrubbing should work. That was a bit of a software development effort which took about three months in total. But once that was done and we had all the integration tested, the actual deployment was basically determined by delivery time of the boxes, and that is true now for the deployment to new sites. Once a box is delivered, it's typically up and running in a couple of days.

Our implementation strategy was to make the solution part of our standard architecture for all data center networks. As of now, we have deployed the technology in 20 data centers around the world. Whenever we start a new site, we immediately put in this technology as well to make sure that we protect our customers on that site. And we try to automate the installation as much as possible so that deployment can be done remotely, from the configuration perspective. That way we don't need to send specialists onsite to a remote data center to get it up and running.

What about the implementation team?

We did the first installation together with the A10 team here in Amsterdam. But all other installations, we've done ourselves, typically with remote hands that have very little knowledge about the specific systems in the data center itself.

The solution is then managed by our team of specialists in our NOC here in Amsterdam. The team that manages it consists of three network engineers who also do other things, of course. They are a part of our network operations team. These three people have developed into specialists for these systems and are, on the one hand, responsible for maintaining them and managing them. But on the other hand, they sometimes get involved when there are specific, large attacks where manual intervention is required to mitigate the attack.

Our experience with the A10 team was very positive, both during the evaluation of the various vendors back in 2015 - A10 was very supportive - as well as during the initial deployments here in Amsterdam where we worked together. They were knowledgeable, responsive, enthusiastic.

What was our ROI?

We have definitely seen return on our investment. If you look at some of the things we can measure, like an 11 percent reduction in support tickets, that can easily be turned into cost savings. Other things, like improved customer satisfaction, are a bit harder to monetize. But for us, we were convinced that within a year, we'd definitely earn back the investment, both in the A10 equipment, as well as developing the end-to-end solution, including the integration with our administration systems.

What's my experience with pricing, setup cost, and licensing?

The financials are always a challenge with this type of technology. That's not really a product-functionality thing but it's the area where we were pushing A10 the most. But compared to the alternatives that we evaluated in 2015, the price-performance of the A10 solution was definitely superior to the other solutions which we evaluated at the time.

The way we did the deal was a combination of the equipment, the license, and a five-year support contract, for all sites. At the time it was a pretty good deal. 

Which other solutions did I evaluate?

We did a lot of analysis in the second half of 2015. We evaluated different technologies and we ended up using A10. We went with it based on the price-performance. We had four systems on the shortlist. We physically tested two of those providers and, at the end of the day, the two came out pretty even from a functionality and performance perspective. But the total cost of ownership of the A10 solution was superior to the other vendor, so we decided to go with A10.

One of the four providers did not want to support a proof of concept test, so we dropped that one right away. We dropped another one after looking at functional specifications which, at the time, were not as good as A10 and the other vendor. We were left with the two that we tested. At the end of the day, the total cost of ownership made the difference.

What other advice do I have?

From our perspective, the technology works well, and it has a low cost to maintain and manage.

One of the biggest lessons for me, in using this solution, was that there are so many smaller attacks going on that we were not aware of and which must have had an impact on the satisfaction of our customers, as well as the satisfaction of their customers. Everybody always talks about the huge attacks, the one- or two-terabit attacks that get into the news. But the fact there is such a huge volume of smaller attacks going, script kiddies, etc., to make other people's lives miserable was, to me, a bit of an eye-opener. That was resolved by deploying the A10 solution.

Availability is very critical to the success of our business. If you look at the customers that we primarily and proactively target, they are customers in the online gaming market, in the advertising-technology/marketing-technology markets, in the Software-as-a-Service and in the managed service providers market. All these companies are borne on the internet and their internet presence is critical to their success, to their existence. So for us, it's of primary importance that we keep their services up and running at all times, even when they are being attacked by cyber-criminals.

As Leaseweb, we have around 18,000 customers using our hosting services. All these customers' services with us are protected by the A10 technology.

In terms of increasing our usage of the solution, whenever we deploy new data center locations, we put A10 in right away. We do have some new locations that will be opening up in the next six months, so we will definitely be using more of these systems and protecting more customers.

I would rate A10 at eight out of ten. What would take it to a ten is the scalability, the ease of scaling up without replacing a box.

Disclosure: IT Central Station contacted the reviewer to collect the review and to validate authenticity. The reviewer was referred by the vendor, but the review is not subject to editing or approval by the vendor.
Add a Comment
Guest
Sign Up with Email