DX Performance Management Reviews

Performance management for capacity monitoring of the Wells Fargo network. 

Performance-wise so far, so good. We have been ramping up close to the limit of the application: four million polled metrics. We are at approximately 3.6 million. So far, it has been doing alright. A couple of hiccups here and there, but overall we are happy with it.

It saves time having it all on one place, so you do not have to jump around from different tools and try to merge the data. It is already done for you.

It does not really affect how the company functions, but it gives us insight into  performance slow downs, and if we need more bandwidth and more larger, heavy-duty equipment.

The integration with NetFlow, the NFA solution, so we can have our flow forensics paired with the SNMP polling performance metrics on the same page. Everybody throws around single pane of glass, but this is really the benefit that we see from it. Have it being able to have everything there in one application UI. 

• Global synchronization errors. Sometimes it just does not finish in time due to the load. 
• When it goes through discovery, or whatever it is doing on the back-end that slows it down, sometimes we get some failures with it. 
• On the NFA ipv6 support, they still have not fully supported it. It is huge for large enterprises. Since we are out of registered IP space, we are moving to ipv6 and we expect our vendors to move with us, and they have not delivered yet. That is a contention point for us. 
• They need to improve on response times for development stuff, bug fixes and enhancement for the turnaround times.

Stability is good. There are definitely some areas where you have patching or a planned outage. There is no solution right now to avoid data gaps. I know they are planning on it and that it is in the development life cycle for the next year or two that they are going to be releasing a solution for it. 

Scalability is great. We would like more because we are almost at the limit or at least the advertised limit for the system. I know they are working to move their four million data pulled metrics up to 10 million. So, that will be a lot better for us, just because we have such a large enterprise. 

Technical support is good and very responsive. We have weekly calls with our account team and the staff. The development piece providing what we are asking for, maybe an enhancement request and stuff like that, is a little slow. However, we like to think that we are the most important customer for CA, but I know there are other customers out there.

Because of security requirements within our infrastructure, the setup of Vertica was horrendous. It was really a lot of work. Vertica is Micro Focus, but it was really hard. 

Upgrades are great. They are very easy.

We were looking at other vendors in the marketplace for NetFlow, and CA just beat everybody else in price, cost, and the size to implement. The infrastructure investment that you have to put into installing the systems. 

We looked at InfoVista. We already have InfoVista at the bank, so we looked at them, but their solution just was not viable for us. 

Most important criteria when selecting a vendor: For Wells Fargo, it is being able to handle the size and scalability. Most vendors out there cannot for large enterprises, because we are pretty much a small Telco in our size. Therefore, being able to scale up to 40,000 to 50,000 network devices is extremely important for horizontal scalability and the layout. 

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.

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.

The auto-certification is valuable to us. It reduces the time consumed to certify a new device, as in the past with CA eHealth.

CA PM is able to monitor more equipment, or only specific requirements, as necessary.

The database should have higher availability in case of an incident.

We have been using the solution for six months.

We haven't encountered any issues with stability. So far, it's been working properly.

It requires higher availability in case of any incidents with any of the servers.

The technical support is excellent.

Previously, we used CA eHealth, which was changed to CA PM.

We would welcome some online demos to be available, or something similar to CA UIM Snap.

CA's support is 24/7 during the entire year and the tools are guaranteed.

We looked at some open code tools. However, they don't have any support or guarantees like CA.

It's a solution which is highly recommended. It will meet and exceed your expectations.

It's a pretty complex product. There are a lot of parts to it.

The biggest selling point is putting the power of creating dashboards and reports into end users' hands. They can create views that are most pertinent to them. That's been the big one for us, and one of the reasons why we went with the product.

From a business standpoint, it's hard to quantify. It's been used in the capacity planning process, as far as our network infrastructure goes. We've been able to utilize data that's been collected by the tool to more accurately predict usages.

The one area where I think it could be improved is the reporting capability. The dashboard features are very function rich, but the reporting feature is probably not as rich as we'd like it to be. There are some limitations in using it.

For example, the number of objects that you can simultaneously show or display in a report is limited. The reporting capability needs to scale up a little bit higher.

Compared to eHealth the reporting is limited. For example the number of graphs in a report is limited to 15. Another limitation is that you cannot copy or duplicate reports.

The whole non-root installation really needs to be clarified a lot more than they did.

We've been using it for just under a year.

We haven't really had any significant stability issues. I'd say that maybe we've had one issue in all that time, where there was data corruption. But we were able to solve that with the help of technical support.

One really needs to make sure that you follow their guidelines and recommendations regarding the database size and performance. They give you certain guidelines that you need to follow regarding the database. For example, given your size environment, this is the size of database you need, and this is how it will perform well. If you want to have good performance, you need to meet these specific requirements.

We have not had an issue with scalability other than in reporting. I would not necessarily say that's a scalability issue of the product, but that's scalability with respect to one of the functions of the product.

I've called technical support on different occasions. Sometimes just for assistance and sometimes to report an issue. They have always been very responsive and quick to call back. I prefer a phone call and I rank them very, very high.

We were using a different CA solution. Specifically, we were using a product called eHealth.

The reason why we migrated was a recommendation from CA. Their focus was going to be on the CA Performance Management product, and less on eHealth. They were trying to migrate users away from eHealth.

The setup was complex. There are a lot of components that had to be installed. The biggest issue that I had with the installation is that it's very much geared to a root installation.

In other words, if you use the root ID to install the product, you're most likely not going to have issues. We just ran into numerous issues because we did not have root access.

While they made some effort in the documentation to describe how to go about doing that, internally it was less than complete. We kept on getting issues that we could not complete because we didn't have root access.

It was a collaborative effort. Through support's help, a lot of time spent on this side, and with support personnel over here and on the server side, we were able to figure out what the issue was.

However, it wasn't a matter of, "Run this script and you're done installing." In no cases were we that fortunate, because we were not doing a root installation.

I wouldn't say it's too expensive. I'd say it's comparable to what you're going to find in the marketplace.

You have to make sure you have an accurate count of the number of devices, and more specifically, managed objects. I'll use that term. It is the devices and the managed objects that you want to put in your CAPM system.

We evaluated other options in the past when we selected eHealth. We did a full evaluation and we went with the CA solution.

After we had only been on it for two or three years, we felt comfortable with CA, their approach, and their strategy for handling network performance management. We just decided to stick with CA and migrate to their next generation, the CAPM.

I would definitely take advantage of the training that CA offers. Certainly take that training, just prior to, or just after you've done the installation. That will put you in a position to really exploit the features of the product.

There's so much you can do with the product and not just the dashboard and the reports that users can generate.

There is more customization that you can do to the product and tweak it to your site's needs.

• Multi-tenant platform
• User-configurable roles & permissions with granularity
• Dashboard-oriented
• Easy enough to customize

Mainly as an MSP configuration, the multi-tenant model helps us to segment the customer and users, as well as providing our operation the capability to oversee all tenants.

Today, tenants are bound to a data collector which requires a lot of networking and resources to configure.

In the context of the multi-service provider and the need to have customer segmentation either for IP overlapping or Security (data protection), multi-tenancy is required. However, in order to do this, the tenant (IP Domain) need to be attach to one Data Collector by Tenant. In the MSP context, it is equal to 1 DC per customer and even for small customer (couple device) this is not too scalable or rather expensive.

I have used the platform since the beginning and in production for at least two years.

Our multi-customer configuration is challenging, as well as to have redundancy over two data centers.

We encountered database issues due to the IOS speed requirement of HPE Vertica, as we are on a VMware environment.

Redundancy over two data centers is one of our challenges, as well as the fact that we use a lot of data collection, even using Docker containers for it.

Support is OK for the standard use of the product but as our configuration is complex, they often have to refer to engineering.

We were previously using eHealth and APG.

As we started from the beginning, it has improved a lot since; once again, our solution remains complex due to our usage.

Implementation was with a vendor team; at the beginning, with CA Service, and then with CA engineering, as we do DevOps.

Before choosing we also evaluated SevONE, APG, and InfoVista.

This is coming along; great progress. The "3 series" looks promising.

The most valuable features to date are the alerts, reports, and integration with other CA products:

• The alerting feature allows us to quickly identify and begin remediation on network devices that are experiencing performance/capacity issues.
• The reporting feature has allowed us to quickly share information with customers, keeping them up to date on critical devices that are being monitored.
• The integration aspect is also important, as it allows us to feed inventory to the system from tools designed to walk our networks, and also pull in flow data for additional troubleshooting in one place.

Utilizing the capacity/performance data alongside the flow we are able to quickly determine the root cause of impact. This allows us to reduce mean time to resolution on network impacting events due to capacity/performance.

There is always room for improvement, which CA continues to do based upon community feedback.

One of the greatest challenges we have encountered are the polling cycles. We currently poll at 5-minute intervals, but at times need to have up-to-the-minute data. Products we have leveraged from other vendors allow us at the click of a button to turn on up-to-the-minute polling for up to an hour. CA Performance Management does not have a similar feature. It may be emulated, however, by setting up a new group for one-minute polling cycles; however, leaving this rate on will impact data storage, so only administrators are provided this feature in our organization.

The solution has been in place for 15 months, and has already provided information helpful in network cleanup.

To date, we have not encountered any stability issues.

To date, we have not encountered any scalability issues.

I have not been required to work directly with technical support. However, based on feedback from the team that does, support often starts with the standard runbook, even after we have described similar troubleshooting steps to them. This results in lost time in resolving issues. Once we have moved beyond the basics, support has been handled well.

Our prior solution was dated and in need of refresh. We switched to the CA solution after going through a proof of concept and rating it against two other competitors.

The initial setup was mostly straightforward, with assistance from onsite technical resources from CA. The majority of issues we encountered were with our corporate requirements on server configurations and not the tools themselves.

Regarding pricing and licensing, keep in mind that bundling will provide the best value. As a result, doing your best to anticipate growth and future needs may save you money in the long run.

We did evaluate two competing options from HPE and Alcatel Lucent.

Take advantage of the professional services and leverage your solution experts from CA. They will help you define the right solution keeping in mind current utilization and future growth.

It is easy to use, and the performance metrics are helping us to know our network better.

It provides metrics to users.

The certificates are not complete. CA needs to provide more support for the product through certificates. They need to discuss this with their product team.

Certs are the drivers to support the devices. Some of the vendors are not fully covered. An example of this is Brocade. Certain types of Brocade used in our company is not in certs published. There are others as well.

I have used this solution for two years.

There was a big problem with HPE Vertica.

There were no scalability issues.

We always use CA.

The setup was straightforward.

The price is too expensive.

We looked at other solutions such as IBM, HPE, etc.

It is important to know what you need.

The benefit is having the one-stop shop where we can integrate all of our CA products into one dashboard. They can use that one dashboard as a portal to click around to wherever they need to go. The dashboard integrates our data from Spectrum and CA PC, like the data collectors.

It has improved efficiency because now there is the one-stop shop where you can view all your data, instead of needing to click around in four different tools. Everybody's looking at the same data. You don't have people using their own tools, or using a different tool. Everybody's using the same data to correlate what is happening.

I would like to see it integrated with UIM. That is the one tool that is not in there. Then I would like to see log analytics, which I know are coming; but are still not included.

Stability is good. It has improved; and we have not had any issues.

At our company, we have a large implementation, and we have not had any issues with scale. We just scale horizontally.

Technical support is pretty good. I would say that we always get to the bottom of the problem. The only reason it is not perfect is that sometimes it takes a couple of days; but we do get to the bottom of it.

I would say the initial setup is complex. CA does make it easy, but you need to have the right architecture for it to be successful. You have to design it correctly from the start.

Currently we have an issue where, if the store goes down, we get 20 tickets spread across different tools. We knew we needed to scale it back to only one actionable item. We needed to use one tool that would encompass everything and scale out, and be able to provide the flexibility to let us narrow it down to what we need.

Our most important criteria when selecting a vendor are scale, stability, and support. We evaluated CA, HPE, Riverbed, and BMC. We went with CA because of our partnership with that vendor. We are most confident in that partnership.

It is a good product that will meet their needs.