What is our primary use case?
Our use cases are for both a government-based entity and an international oil and gas entity, and so our use cases flow for security across both domains. Very similar threat cases are created, but they're targeted specifically for the client's operating environment, including standard access control, endpoint control, and things like that. The use cases themselves vary from active directory exploits to endpoint exploits. We use it for real-time alerting, so we run an alert-centric model that we partnered with a log service and we do a discovery centric model on the back end, so we have a hybrid.
How has it helped my organization?
Some of the benefits of using this solution are rapid correlation and near-time response on alerts.
What is most valuable?
The UBA features and, again, the correlation engine is nearly bulletproof. Once you have it dialed in, it provides accurate near-time responses as things are coming in to correlate and identify.
What needs improvement?
ArcSight is incredibly complex when configuring and deploying, and if your organization doesn't know what they want and what they need, ArcSight will be a challenge for them, but if they absolutely are skilled and know what they want and know what their end goal is, ArcSight is brilliant.
I would like to see more dynamic reporting.
It could look a little newer. The interface looks a little dated, but otherwise it's fine.
For how long have I used the solution?
I have been using this solution for about 10 years.
What do I think about the stability of the solution?
This solution is rock solid. The stability is untouchable. If you've scaled the solution correctly, the performance is brilliant. I rate the stability as a ten out of ten.
What do I think about the scalability of the solution?
If you have employees with knowledge, the solution scales vertically and horizontally. You can just sit there and say, "I'm ingesting this many more log sources, I need this much more processing, and this much more storage," and you can just drop and place. It's easy to scale because we're using clusters and blade servers, so we just slap in a new blade, add it for the cluster, and now it's scaled.
We have SAN storage, all high-speed disks for the 90-day storage, and then mechanical disks for the 188-day storage, and then you go to cold storage after that, which is disk as well, and it's a SAN, so we just swap in new JBODS every time we need more storage.
There are about 40 people using this solution from an analyst point of view within my company. We have threat hunters, endpoint protection research, vulnerability research, and security analysts on levels one through three. Pretty much anyone that operates in a security operation center is using it. We run three shifts, 24 by seven. Each shift is an eight-man team of analysts, and then we have endpoint protection subject matter experts, forensic subject matter experts, and threat hunting subject matter experts all using it Monday through Friday. At any point during the day, there's about 40 active users on the solution. All in all, I think we have 150 named accounts inside of it, but usage, there's about 40 at any one time on it.
I rate the scalability as an eight out of ten.
How are customer service and support?
Micro Focus is less helpful than HP Enterprise was. If I were to rate HP Enterprise when they owned it, I would give them a nine. I'm going to give Micro Focus a six because, after we started operating it, they would provide us professional services or engineers from less desirable skillsets and locations. Being in the Middle East, it's difficult to get quality people, so we ended up getting a lot of people from South Asia and Central Asia that just weren't as good as people we had before coming out of Europe or North America.
Which solution did I use previously and why did I switch?
We migrated off of QRadar to ArcSight. We switched for better performance, capability, and usability.
How was the initial setup?
The setup can be very complex, depending on the size of the organization. Our organization is huge. We have a vast ArcSight infrastructure, high availability, and multi-noded clusters. Our usage is very unique in that way, and it's very advanced and complex. Most organizations probably won't have the level that we have. The setup is a ten out of ten in functionality. On configuration and getting it set up, it's a one because, again, it requires some very specialized knowledge.
What about the implementation team?
We used two architects and three engineers on our team for deployment, and the team from ArcSight had nearly an equivalent amount. We handled deployment in-house, and we fully deployed it enterprise-wide in about six months. We had HP ArcSight certified engineers and architects, and then we sent a handful of our own engineers to HP so they could become fully ArcSight certified. Their engineers and our certified engineers then worked hand in hand in kind of a mentor, mentee relationship to ensure that our team had full knowledge and capability going forward.
The first step of deployment was scoping and sizing for five-year growth, based on what we were currently running in the older product, which was QRadar, and so then once we determined what size infrastructure we needed, we deployed that infrastructure. That took about a month. From there, we then on-sourced non-critical assets for testing and piloting. Once we had that done, we deployed the agents for the use to our SOC, and then we ran both systems in parallel to make sure that use cases reported over correctly, and they were all fine-tuned.
Once we had them working on our test samples, we then did a rapid deployment across the entire environment. We ingested everything from the old system into the new one to the log collector. Once all the old logs were in there, we then switched over to real-time and transferred the real-time logging from the old system to ArcSight, and then that system was live. We did one after the other, and that's what took the six-month window, because after about a three-month deployment of getting all 35,000 log sources ingested and up and running, it took about another three months to do the rest.
What was our ROI?
The ability to detect and respond to an event within 60 seconds is priceless to us because a detected for each event becomes escalation, and objectives on target add about 60 seconds or more. Having ArcSight provide us these alerts within 15 to 30 seconds of something actively happening, where our analysts can then begin to engage, isolate, and mitigate, that's priceless.
What's my experience with pricing, setup cost, and licensing?
The solution is super expensive. I think we're paying about 1.2 million a year. We have a full, perpetual, all-we-can-eat license, which is why we pay that, because at any moment we may onboard entirely new infrastructures.
The problem with QRadar is that you're paying EPS licensing, and the moment you exceed your licensing, QRadar stops processing anything above that license count, and that becomes an immediate security event. With our license with ArcSight, we don't have that problem. You pay for all-you-can-eat and we can give it all-we-can-eat, as long as the hardware that we have running it can support it. The hardware is ours, so we can just grow that as we go.
Part of our yearly price model is also including extra blade servers and disks, so as we need them, we can just slap them in, and boom, we expand the cluster by three or four more processing nodes on the blade server, and now we can double the amount of queries and searches.
At our organization size and license model, I think the price is average to what anyone else would charge us. I would rate the price as a five out of ten because obviously we would like to be cheaper, but we know what we're getting for that money, so we're okay with that. We're neutral about the price.
What other advice do I have?
My advice would be to ensure that you understand what your goals for the product are, and ensure that you have the correct engineering knowledge to implement those goals.
I rate the solution as a nine out of ten.
Which deployment model are you using for this solution?
On-premises