DX Performance Management Reviews

I'm using this solution for topology and dashboards as well as historic trends. I'm a solutions architect and customer of DX Performance Management. 

Dashboards are the most valuable feature of this product as well as integration with other solutions. It helps with issues in the network and replacing devices. 

The strategy and the policies could be improved and I think monitoring could also be improved.

I've been using this solution for 10 years. 

It allows us to have a quick, upfront view of our multiple data centers. We have multiple connections to hundreds of customers; and it provides us with a quick dashboard look at issues we might be having; bottlenecks we might be having; and things that might go wrong in our network without us needing to dig through endless screens of issues to get to the root of the problem.

It makes things faster when we have our network operations center. They can look at their dashboards that we've created for them, and they can see problems immediately. We can catch a problem before it really starts impacting the customer. It enables our network engineers to look deeper into the source of the problem after alerts are raised. So it's quick, it's fast, and it gets things done. We can get past a problem before it becomes a real problem.

What's missing is stuff that they're already putting on the roadmap: the quick dashboards that they're building; the ability to create dashboard quickly; for example, with ADA and NFA, has been pretty amazing.

They are starting to support additional products and every time we come out here, and every time there is a product release, something new is added to it. Honestly, CA should just keep doing what they're doing. I'm sure there's going to be stuff out there we haven't thought of; and I hope they figure it out first.

Stability has never been an issue. That's one of the things that's so great about CA products is that we've never had that kind of issue. We don’t have servers crashing on us for CA products. We're not dealing with issues with connections or whatever. CA just works. There are no stability problems whatsoever. So it's fantastic.

Things work. I get up in the morning, I open the console, and it's there. I don't have to worry about it.

We have multiple large data centers and we've never had any issue getting those data centers hooked into CA. If we have to connect to something with our ADA or NFA products, we find it: it's there. We don't have any problems with size or scalability. It's quick, and if we need to bump it up a little bit, we can do that as well.

We use technical support all the time. In fact, they just resolved an issue for us this week with a name tag for IP SLA products in Cisco. The nice thing about them is that they're so quick and responsive. When we have a problem, they will get on it, and they will find us an answer, usually quickly. They come back and fix problems and bugs, so that it worked great. Their technical support has always been fantastic.

Actually, I was not involved in the initial setup. I inherited an old mature system. It seems to be straightforward. I worked with a guy who did set the product up initially, and he had no issues with it. But again, I'm now working on a mature system, so I haven't done that yet.

Get CA involved up front. Let them help you. CA has done this for years. Let them help you with the things that you might not even realize that you don't know. Let them help implement it with you because they're going to know some of the bumps and hassles and pitfalls that are out there. That's what my partner did before I came onboard; and he said it was as smooth as pie. So, that would be my advice.

We are a service provider and we have a lot of in-house servers using DX Performance Management. On the customer's network, we provide events and performance data, and we can control the network, such as the performance, and give service to our customers.

The most valuable feature of DX Performance Management is the quick way we can collect data without putting much effort. They have a lot of customization and features with the solution internally and out of the box. We can discover the device, and start collecting information. For example, if the device is available, we can receive the CPU and memory utilization, and the interface optimization is simple and very fast to implement.

The solution could improve the database model. It is also not very good because they have a lot of interconnections with other data sources, such as NFA and VNA.

The synchronization when we have tenants could improve. For example, the multi-tenancy synchronization process with the other data sources could improve.

In a feature release, the solution should have templates. It does not have tenant templates that we need, and there are a lot of features that we would like to have but do not. We are in touch with Broadcom and are asking them to implement these features in a future version.

I have been using DX Performance Management for more than five years.

The solution is stable but has some issues with synchronization with DX Spectrum that we use.

I rate the stability of DX Performance Management a four out of five.

DX Performance Management is a highly stable solution.

We have a lot of customers that use this solution because we are an MSP. We have over 1,000 users and 80 servers.

The support quality depends on the issue that is presented to the support.

There are issues that we have, but the support is very good. Other times when we need to involve engineering and the second line of support it is not good. Additionally, parts of the support come from Europe, the United States, and others from India. It depends on the issue we are facing to where the support will be received. For example, web access to the development team is very difficult.

I rate the support from DX Performance Management a four out of five.

Positive

We use DX Spectrum for full management and DX Performance Management to correct performance data.

The setup of DX Performance Management is not difficult but when we did the first installation, we had some support from Broadcom.

It is difficult to say how much time it took to do the implementation because we did it in multiple steps and multiple phases. Phase one was more than five years ago, the installation was done in one to two days. We did the integration with multiple servers it the whole process took approximately two to three months.

Our first step in our implementation strategy was revised because we are an MSP and we need to involve the security and the data center teams to revise where we need to put the servers from the customers who can connect the networks. It took us three months for the first phase of the project and the full completion of the project took approximately one year. There were a lot of servers and customization involved.

I rate the initial setup of DX Performance Management a three out of five.

We had some support from Broadcom for the implementation of the solution. We had two people from our team and one person from the Broadcom team that participated. The assistance from Broadcom was once per week.

We purchase a license to use the solution every three years and we have multiple contracts. They have been increasing the price annually. We have a baseline for three years and then we need to pay the difference from the baseline, and we are in the process of negotiating a new price for all the licenses we need. The price overall is affordable.

Overall the solution is good.

I rate DX Performance Management a seven out of ten.

Probably the most valuable feature is the integration that the tool provides for us. We're able to present data from multiple back-end collection sources into a single dashboard for the users. So they don't have to go to multiple locations to get data about a particular item, or device.

Stability has been fantastic so far. We've got quite a few different options available to us for business continuity, things like that, and the just inherent stability has been fairly impressive so far.

Scalability is the reason we bought the product to begin with. It was designed from the ground up for carrier-grade services, and we are in effect a MSP ourselves. So we were really interested in looking at something to be able to handle the multi-tenancy and scale as large as possible. This was the only solution that we really considered at that level.

Everything has been very responsive to us so far. I've been impressed with the followup that the engineers have as well. Even after we have fixed whatever issue has come up, it's not uncommon to get a followup email a week or so later just checking to make sure everything is still all right, and I appreciate that.

Previously we were using multiple products, multiple individual point solutions, and it was getting to the point where it was really difficult to maintain every one of those individual products. That's the reason we looked at Performance Management, because it gave us one location, one place to do all of group administration, and to build dashboards, and device administration, inventory counts, things like that. So it really reduced our overall administrative overhead.

At this point I would give it an eight out of 10, but that's only because I know of a lot of really cool new stuff that's coming down the line, that's not available yet. So that number is subject to change.

Our primary use case for the product was to actually monitor the devices in our network. To actually help alert on problems and issues, then do some type of self-healing for the issue. It was also there to find a way to build trends and look at our network to see what we look like now, and where we may have to plan going into the future.

The tool does pretty well. It is pretty new to us. We are still learning it. It is a broad value. People were a little hesitant when they started using the product. Now, it has become a vital tool for our use in the company, from a network perspective. 

Since the implementation of the tool, it has cut down on probably 60% of our outages and letting us know what is going on. That little component has created significant improvement for us that we were able to put in place. 

There is another component of the tool called Network Flow Analysis. It gives us the ability to troubleshoot issues which do not appear right away. It was actually an added benefit to the tool for us.

Its ability to probe the POLAR Network and tell us about a device. That is the most valuable. It is important we are able to alert based on what we find. It has become useful, because we can interact it with another tool, like Splunk, to actually do other parts of alerting.

Since the direction is more cloud-oriented, I would like to see what modules or functions they are putting out there to say, "Hey, this is what we are doing to monitor your cloud or help monitor that issue." As we move more towards the cloud, we have to monitor what in that space can I do with the CA PC tool in that environment.

There are versions that are better. One of my concerns was it did not have a built-in HA component for availability, which was somewhat of a challenge for us. Because as we became more dependent on the tool, we had to find a way for it to be up all the time. So stability, I give it about a 85%, but we are getting there.

They have actually announced other versions that we need to get to. The challenge we find is, because this tool gathers so much data, which makes it hard to convert over. To upgrade it, you really need the solution backed up. That is the issue that we are facing right now.

I am happy with the scalability for right now. It is better in the later versions.

The company is 25,000 employees. Our department consists of two groups that work together. We are broken up into a monitoring group and a network group, both use the tool, probably about 20 of us. Maybe a little more that use the tool to get what we need out of it.

Customer service: I am quite comfortable with CA and love the people that we talk to about it. They are responsive working with us and made us feel like they really care about the product. I get that from the sales perspective, my representative. I would not trade him in for the world. He has done a really good job. I am really happy with him.

Technical support: I would give them about a "B" right now, because they do not respond as quickly as we like. They do get around to us with some issues. It seems like we escalate more than I would like to. If anything, they should look at how the tier support goes in place. 

If I call and it is critical, then I have to wait for someone to call me back with an engineer, I am not too happy about it. Because you only call when you need it, so if I need you, I need you right then. If it was some other issue, such as searching for knowledge, I can understand opening up a ticket.

We normally would not call unless we are trying to do something with the tool, and it is not performing the way that it is suppose to, and we are trying to find out why. Normally, I call because I need the tool up and working, then I need to know, "Why?" However, first of all, I need the tool up. 

The issue: There are sometimes they have to go and do research and get somebody to find out why the tool is down for a minute. That could be somewhat of a challenge for us, because we have people that are looking for the tool.

The initial setup, I wouldn't say was totally complex. I would give it on a scale of one to 10, it is about a six for the initial setup. Once you have the information, I did not think it was that difficult. The challenge was to find out what platform they were running on.

I come from a company that does both Windows and Linux. Of course when we asked them, which environment the tool should run on, they said they support both. They still support both. What I would like to know is, what is the most stable environment for the tool? 

That is where I am looking for them to say. I know no one wants to choose one environment over the other. I do not care what you say. I just need to know your recommended way from experience, which environment has less issues on it. That makes a difference.

We are using Linux, and I am glad we did. I think it would have been a disaster the other way.

We looked at SolarWinds. We looked Riverbed. I think we just looked at CA. We did not go looking at a lot of them. CA also came to us, because we had someone working on the other side to say, "Hey, we already had them in-house, take a look at a tool they had in place."

That helped them, because they were there. Once they could provide the tool that we wanted, and they were able to go through our use cases along with other vendors, that is how they got selected. They fitted with everything that we needed.

Kick the tires and get under the hood. Provide your use cases upfront, and tell them exactly what you need. Once you do that, it is fine.

They will provide you with what you want, and meet your criteria. Lay out everything that you need. Look at the sales person that you are dealing with (the sales engineer). See how knowledgeable they are about their product, and see if it does what they say it will do, and that it is available right now. Do not tell me something is here, then, "Oh, that will be available six months from now." That does me no good. Unless you tell me up front, "We're gonna have this in the future," then I will know.

We are still learning it, because there are still features that we need to implement. It is a slow learning process. Right away when you get a tool, you want it to be implemented to actually address the issue that you have at that time. Now, we are looking at it to see what else it can do for us. 

From when we first got the tool, it was there just to look and discover devices, and tell us if they were just available or not. Now, we are looking at it to say, "Well, we can trend on certain things." 

Most important criteria when selecting a vendor: 

1. Be able to meet my requirements.
2. Stability
3. Price. That makes a difference.

Our primary use case is for network operations. We use it for customer information, as well as for information for our operations team. We are a service provider and we give our customers access to the Performance Center interface because it gives them information about the health of their network in different locations.

For example, for a large customer with, say, 100 locations, each of the customer's employees can see the network health in his area. That means that they provide them, for example, the WAN bandwidth and DNS requests and network-related metrics.

The first-line operators, who are sitting in front of the monitoring consoles can easily see if a problem is related to a customer location or to the complete customer network. It's very easy to see because you have the top-20 views of metrics. It can highlight where the problem is. It's very easy for them to see.

Also, the capability where not only the traditional SNMP information is captured but also the netflow data; who is consuming the data on the WAN, and voice-related information, is helpful. The voice quality makes it very easy for first- and second-line operators to see where the issue is, and who is impacted.

For us, the multi-tenancy is very important, of course. 

The scalability is also important because we have customers in the system with only a small number of devices, say, 50, but we also have customers with more than 10,000 devices. The scalability is very good, amazing.

There have been a lot of changes which were very good in the last few years of development. One of them is that they brought together Spectrum and the performance data, since Performance Management 3.6. They also have the alarm information in Performance Center which is very helpful to have in one tool.

Every product has room for improvement. It could be a little easier to integrate new metrics.

In the beginning, there were some issues with the Vertica database they were using. There were some issues with stability and memory leaks, more related to memory leaks, but this was fixed in the past. At the moment, it's all very stable.

Especially for the issue with the Vertica database, we contacted the support, of course, and it was okay, it was good. It was not negative. The response is typically very good, especially in this area. Of course, because it was the database behind the issue - it's an HPE database -
it took a little time until it was fixed. But that had nothing to do with the support guys.

Before this solution, CA didn't have an equivalent solution and we used open-source solutions. But the capabilities weren't there in the open-source solutions. With this solution, we reduced about 15 performance management to only five machines. The open-source tools were not able to deliver this performance.

It was quite easy. The setup was done in a couple of hours.

Not in the first version, but since, the integration with Spectrum has been quite good. It's more or less fully automated, which means that we get new devices into the network monitoring. It's also synched into the Performance Center. Similarly, if there are new routers sending natural data, it's automatically be coming in, which means that it reduces our administration costs.

With the capabilities of the open API and the web service interface, we're able to automate some other things. For example, we have one customer, with more than 1,000 devices, and there are a lot of changes in infrastructure every day. We automate the menus and the groupings of the devices. We can automate that via RESTful web services. That capability is quite good so in this case, it's all more or less automated.

The deployment, for new customers, takes less than a day. Typically, the products are not the problem. Usually the problems exist in the network by itself: They don't have access to the devices or the like. In the case that the infrastructure is ready, so that you can run your setup, it's quite easy. After a couple of hours, you see the first data.

My colleague does the deployment. Just one person does it.

I like the licensing model because it's device-based. It makes it easier for us, as a managed service provider, to bill our customers. We are paid per device and the license per device makes it easy for us.

You have to think about what you want to capture. You can capture a lot of things, but you have to make sure you capture only the things you need. We have had situations where a customer has said, "We need this, this, and this," - hundreds of parameters. But that would mean their infrastructure needs are growing massively. You have to think about the metrics you really want to capture. You can capture everything, but it doesn't make sense because then you pay a lot for hardware.

For our customer with 10,000 devices, there are more than 700 registered users. That doesn't mean 700 users at the same time, but we see 700 registered users using the tool regularly. On the other systems, there are around 200 users, including networking guys, first-line guys, second-line guys, etc.

The maintenance for Performance Management is quite easy. You have to do one to two releases a year. Testing it on a test environment takes one to two days. And then we'd need about five to ten days a year for maintenance.

We aren't using it as extensively as we could. It could be used more. It's a process to bring it to customers and show them the improvements, especially in the last versions.

I would rate the solution a nine out of ten. It could be a little bit easier in the administration or in the creating of new metrics. The really deep-dive administration could be a little bit easier.

We use the Performance Center to do performance monitoring. It is used throughout the organization by both our engineers, our support staff, and our troubleshooters.

When devices are having performance issues, it proactively build dashboards, which allow us to go and do health checks, and resolve problems before they become an issue.

The ability to quickly do drag and drop customized reports for dashboards.

I would like to see open integration with tools like Splunk, ExtraHop, and NetMRI.

There are some areas in the technology right now, like with VMs, where we are lacking with our abilities to get inside the VM to monitor traffic within the machine.

It has gotten a lot better over the last five years. When it switched from MPC to Performance Center, there were some issues, but it has greatly improved.

I have mixed feeling about the scalability. I feel like there are things which are being put into UIM right now that are not being included in Performance Center that we need to see in Performance Center. We are kind of being driven to buy UIM, and I can't justify it.

There are just some charts and views that we can't build in Performance Center right now.

We have had a couple of instances with one support person that we've had some issues with, but overall I would say it is very good.

We brought a CA person on site to help us with the initial setup. They were helpful. They got us up and running.

Have an open mind. Look at what it can do for your folks. How fast and how easy it is for them to learn to navigate through the dashboards, and what you can do with the dashboards.

The integration to the other products that we use: NFA, we use ADA, we use Spectrum; and CAPM integrates all those into a single pane of glass, for performance. CAPC is good for developing reports around those multiple data sources and giving us a single vision on what's going on in our environment.

In the past we've had multiple monitoring tools and products being used simultaneously. The goal of the product - and we're still working on eliminating some of the overlap that the other products have - is to allow us to have a single monitoring system, or as close to a single monitoring system as possible on network performance. That way, the network engineers don't have so many places to look for performance data. Prior to using the CA tool suite, we probably had - and I would hate to say this - but we probably had a hundred monitoring tools.

Some of the individual report views, the way some of the columns sort, there's room for improvement in giving us more flexibility in being able to sort reports based, for example, on what columns the metrics fall under.

I would say for the most part, most of the improvements I would identify would be on individual reports that are produced by PM, that we instrument to be produced by PM. They may not necessarily do everything we want them to do.

We've had CAPM deployed a little over two years. I have experience with CA's NetQoS NPC for about five years prior to that. CAPC basically replaced NPC.

Not really stability issues. We've run into bugs but I wouldn't call those stability issues with PM, specifically. We have run into software bugs where certain features weren't working right, in the past, that CA would have to address. I don't think that would fit under stability. Stability's like crashing.

Now, the one thing I will say. When our CA performance management system is on a Linux system, when that Linux system needs to be updated with patches - and this is the server itself - of course PM has to stop. All of our monitoring stops while that patching is going on for those services, while they restart those servers, but I don't know that that's a CA thing. While that reboot is going on for those servers, for those patches to take place, of course the system has to restart.

No issues with PM. We use other products like NFA and ADA. We've had scalability issues in those areas, but with PM we have not. I don't think we've run into a situation where we have too much. We actually designed our system around a million interfaces to monitor, and I think we're probably less than half a million at this point.

I would probably give them an eight out of 10.

We switched because there were multiple other solutions, and we were already using CA Spectrum, so the natural progression was to go to CA and use their performance management tool, also known as CAPC, Performance Center. We used multiple tools. We had NetCool out there, we had SolarWinds out there, CiscoWorks, numerous tools.

For us, I would say it was complex. But I think a lot of customers that use the tool rely heavily on the out-of-the-box reporting that CA produces with the tool naturally. For those types of deployments, it might go easier. For our needs, we highly instrumented reports. We have built literally hundreds and hundreds of individual reports ourselves, for our own needs and, I would say that that comes with a level of complexity to accomplish.

I was not on board at the time that they were doing of the selection of this tool set, with Anthem. I'm aware that they were also looking at NetCool as another option, and I am not sure why CA was ultimately selected. I also don't have any information about pricing. I'm not sure what they spent on it.

On the positive, if you can rely on the out-of-the-box reports, it should be a fairly straightforward deployment. If you're looking at instrumenting your own views and stuff, it gets more complex the more views you want, the more complex your views, and the more you want to instrument the tool for your environment, versus using the out-of-the-box solution. My advice would be to understand how you're going to implement this and what you're expecting from it.

You're going to go one of a couple different ways. You're either going to instrument it to satisfy some specific needs that you have, or you're going to depend heavily on the out-of-the-box reports that only do instrumentation in a few examples that you find you need. When they demonstrate it to you, you should ask those questions about the differences.

In follow-up to my rating it an eight out of 10 overall, I've used a few other products. There aren't a whole lot of other products I could even give a five to, to be honest. I've used SolarWinds, I've been exposed to NetCool. I've heard horror stories about HPE OpenView. I would give it an eight, which isn't perfect but it's high on my scale.