IBM SevOne Network Performance Management (NPM) Reviews

We use SevOne to manage about 10,000 network devices on our system. We monitor those devices with all the performance data, run reports, and see alerts. We have a manager of managers that sits above SevOne that actually displays all of our alerts, does some correlation and other things. We also provide some maps and reporting.

SevOne also enables us to detect network performance issues faster and before they impact end-users. We were monitoring the load balancers on our backstage passes for access to the network. And we can see, it went from around 3% to around 75% over a couple-of-week period where they had to send in all the remote access and change everything. SevOne really did a number for us, during the pandemic, of isolating which load balances were overloaded with users working from home. So that right there, was worth its weight in gold, because the management created all these reports for load balancers, for access for remote workers, and that's all they focused on, for a couple of months. So that was nice.

It has saved at least 50% because if we're just using ping and a couple of other tools, you can't really see that, all these devices went down at the same time, that segment, or that peer.

Data Insight reporting tool is the most valuable feature. They came up with it a couple of years ago. The most pleasing factor is the dark theme. You don't have a white background. It has templates that you can create for all kinds of reports that you can hit on the fly. It has a much better printing of the reports. If you want to send PDFs to people, the reports are actually decent. Whereas for years, the old architecture of the PDFs was rubbish and even our customers said, "We have to manipulate your PDFs because they all have bad margin breaks. SevOne fixed that a couple of years ago with the new Data Insight. It's fantastic. I would say the reporting of the new Data Insight is my favorite feature. 

We also have the Wifi Controller feature and we're starting to turn that up. That's going to be nice because we're going to be able to monitor wifi. Our group used to monitor wifi, about 10 years ago, maybe even longer, and then they took it away and gave it to Cisco Prime LAN. And they come to find out that Cisco Prime wasn't monitoring it as well as they thought. So we got some quotes from SevOne for a wifi solution, and now we're implementing that. We're excited about the wifi solution.

We also use NetFlow and Databus. It's not that new, maybe five years old. But everybody's starting to get on board where we just send our raw data to scientists. They correlate all the data into how they want to report on it. Those are a few of the new things that we like to use.

I would rate the comprehensiveness of SevOne's collection of network performance and flow data a ten out of ten. I've used Concord and eHealth before this. I used HP OpenView for 15 years. Right now, SevOne is top-notch for me because it's an all-in-one package, and it's easy for the operator to learn. If I can learn it, anybody can learn it. But it has a lot of features underneath that. I am one of the admins, but we have some really top-notch programmers that go in and get that in-depth data. I operate as an admin, I help people out, create policies, and everything. But when it comes to the in-depth stuff, I leave that to the scripters. I'd rather just click on the GUIs and let somebody else scrub through the comments.

It's extremely important that SevOne's collection abilities cover multiple vendors' equipment. We have F5 Firewalls, Palo Alto load-balancers, intrusion protection devices, ClearPass servers, Aruba, we got it all. SevOne has a good process. We also like the certification where we get the MIBs and the OIDs from the customer or the vendor. And they say, "We'd like to monitor this CPU key performance indicator." Or "These HC octets and the interfaces. If it's above 80% we want an alert."

With the vendors, we just take a new vendor like Aruba, they'll want to monitor the fan speed or whatever, we'll take that OID and send it to SevOne. Their certification team is top-notch. They have a 10-day turnaround, but for us, they always provide it quicker. We tell the customer 10 days but we sometimes tell the customer too, that they're always quicker. And they always are.

The process is easy. As long as the homework is done ahead of time, either by us or the vendor, we just provide SevOne with the OIDs, they provide us with a file, and we import it into SevOne. We apply it to the right vendor and all our key performance indicators are there. It's wonderful.

We're also just starting to monitor software-defined and streaming telemetry-based networks in our environment. We got a new manager and he's been pushing it. He loves SevOne. We use Data Bus, NetFlow, and we're doing the telemetry stuff. I don't really understand it, but we're working with some scientists on ride controls, to send them that data. When they started doing this, I told them "You better get some sharp people down here." And they did. 

The manager is a great manager. He's holding everybody's hands to the fire, and I got a bunch of burn marks on my hands. But we're getting progress. SevOne was great, but we weren't taking it to the next level. And other people were coming up with other tools, saying "This tool does this." And we said, "Well, SevOne does that, if you want us to do a proof of concept." So we've been doing all these proof of concepts.

In the old days, reports had nice baselines and stuff that we could use for deviations. With the new Data Insight reporting tool, now we have percentiles that we could have in the old ones, but when you had a reporting tool that wasn't that good, you're not real excited about baselines and stuff.

With Data Insight, we can see baselines and deviations. We can decide how many deviations we want to view. We can do percentiles. We can do time over time, and the graphing in which you can separate the graphs. Data Insight is a game-changer for reporting. 

You can look at the reports and it's just a picture, so your brain can say, "Whoa, that's out of normal. There's the baseline and there's somebody making a backup in the middle of the day or something." So, the out-of-the-box reporting is very nice. Every time they upgrade us, they upgrade Data Insight and they add more templates that their team has decided that the crews could use out there. They're great. I always see the new templates and I just copy it all over to my environment and change the names so people don't see.

The dashboards are fantastic. I don't use them as much as I should. I just started creating some. I'm doing it in the new Data Insights. You can customize it to your customers. We don't do much of that because we don't have a big enough crew to manage all the users out there, there are hundreds of users. And if we had to be their reporting gurus, we'd be hung up all day long, just clicking on reports for people. 

I love the dashboards because you can put it all in the front. You can have heat maps on the CPU. If you want it to have a dashboard for all of F5 you could just have the dashboard for F5 and say, "Hey, we're having CPU problems. I just want a heat map. Show me something red that I can click on and go troubleshoot." It's so nice.

There are a lot of pain points. My main problem is that we don't have a high availability system. There are 20 peers. We're going to lose the end-of-life appliances that are old. If we lose a peer and it doesn't come back, we lose all that data. The reason we don't have high availability is because it's double the charge.

I wish there was some way that we could just get a snapshot of our system so that if one of our peers failed, we could go through the process and get it back to where it was. If we built another peer, and it took us four days to build another peer and get all the firewall rules and everything it would be nice when it came back if we had a snapshot that said, "Hey, peer two, that died." Then can we just slap all that data onto the new peer two and have all that historical data, as opposed to just importing it new, and it wouldn't have any data from the past. That's kind of a pie in the sky thing. But I would like some kind of backup system.

I've been using it for about eight years now, which is actually a long time. Usually, our applications come and go. 

The stability is dynamite. We are having some issues in our VM world, where we don't have visibility to our peers that are out in the VM world. Sometimes our teams might get a peer locked up or whatever, but it's never SevOne's problem. When we had our appliances, it was rock solid. There were no issues with SevOne. You had a disk array and if you had a disk that went bad, you just ordered the disk and dispatched somebody out. I'd give them a positive as far as stability.

Scalability seems to be incredible. We're adding peers one after another. We got the wifi solution and then we just added four new peers, two on the east coast, two on the west coast of the United States. We just order more peers and get them built. SevOne sends us the OVA files. We install it, we open up a case of SevOne. They help us bring it into the cluster. And boom, we've got another whole peer ready for another 1000, 2000 devices. So its expandability is very nice, much better than OpenView and the other things I worked on.

We previously used HP OpenView. That was my thing. I liked it because of the maps, you could have all kinds of cartoons and stuff in the background. That was fun for the graphic artist people. SevOne just blew HP OpenView out of the water.

We had four servers and around 10,000 devices out there and we just couldn't handle it, it was just too much for HP OpenView. HP OpenView stagnated because I used it for about 15 years, and the last five years it looked like it was dying on the vine, with the support and stuff. They changed systems and our people in charge of budgeting and projects, decided not to go the route that HP suggested and went the SevOne route, which I'm glad they did.

I was the sidekick for the setup but it seemed to be pretty easy. I had installed, from setup, HP OpenView systems with four D80 servers around the world. The SevOne environment was pretty good. We were small at the beginning.

Without the planning and everything, just when we got the devices and turned them up, it took around a week or two. We were in our own little lab, testing.

We had a database and we were taking Cisco devices first. Once we had all key indicators identified that they wanted to monitor, we did it. Then we slowly brought in each vendor with the certified files and checked them as we imported them. It was a good plan.

We've used SevOne any chance we can get. We call them in all the time. They have a really tight relationship with my boss. They bring them in whenever there are questions on anything. And their support team is fantastic. We open up calls and get our tickets taken care of nicely.

SevOne is definitely earning its money because different departments are requesting SevOne monitoring for certain situations. And it's extra-billing, of course. I never see any of it, I just see the devices and we add them and we charge them. So they're bringing in money. 

They're getting their money back.

They were constantly looking at other products. I don't look at them. I don't even have time to think about other products. They looked at NerveCenter but NerveCenter is different. My customer is constantly looking for other replacements that are cheaper. Everybody's looking at their budget and asking "How can we get cheaper?" 

At one time they suggested ThousandEyes. It's much cheaper and easier. Well, they had ThousandEyes monitoring a little section of their network and they realized that there's no way ThousandEyes can do it. It's just too big of a network. ThousandEyes can do little stuff but overall, I work on changes all the time and I do my SevOne stuff, and the guy does his ThousandEyes stuff and his stuff is not quite right. 

My advice would be to read the PDFs they have and then look at the videos on YouTube. That's what I do. I'm not a voracious reader, but I go to YouTube a lot. 

I would rate SevOne a nine out of ten.

• Quick response on reports
• User interface
• Scalability of the monitoring
• Reliability

Given the ease of access to the information in a few clicks, the user base of the product has increased tremendously. As the word of mouth spread, the increased reachability of this for the performance reporting space within our organisation increased.

Keeping up to date with market trends, new vendors and with network vendors’ product support.

More than three years.

Our installation is a very large scale one. We initially had issues getting the product to accommodate the number of network devices to monitor. After a set of fine tune steps, we did not need to turn this part back on.

We are one of SevOne's large appliance base customers, and we have not had any issues so far.

8/10

The predecessor was not up to the race! Over the years of it use, it failed to meet the expectance organically. Hence the need to look for a product with better returns in many aspects.

Straightforward I must say… Just get the appliance set up as a vanilla installation in the management systems. It’s sort of plug-and-play. It covered 75% of the network devices and servers farms out of box from day one.

SevOne has a decent pricing model, so far, for production & solutions on the shelf. Given the “object” licensing concept, it simplifies the pricing model in many ways, and gives flexibility for customer to decide.

Thoroughly evaluate what needs to be monitored from SevOne, you’ll look at it as a different ball game for the cost of what needs to be monitored.

Solarwinds – Orion, Nimsoft, IBM TNPM, Netcracker, HP Performance Insight, OPNET Net One, Previsor, TBD Fusion (Vital Suite), Watch3Net, ScienceLogic, Cisco – Prime, Network Instrument – Observer, ZenPM - SysMech.

Clearly outline what needs to be monitored from SevOne and what you aim for your organization. I don’t recommend monitoring everything.

Please engage the SevOne SME’s. It will make the implementation easier in many aspects.

We are using this solution for monitoring the network for performance and availability. We have about 25 SevOne peers that are monitoring almost 8,000 devices. These devices include routers, switches, firewalls, etc.

On any outage, SevOne is pretty quick to send an alert. We've got an operations center that consumes the alert and sends it to the device owners so that they can minimize the time of impact of that alert. Such outages happen at least once a month, and whenever there is a real outage, SevOne is the one to detect it.

The comprehensiveness of SevOne's collection of network performance and flow data is very good. For NetFlow, I would rate it a 10 out of 10 because it collects everything that NetFlow delivers. You can also customize the reports to show only what you'd like to see or what your customers would like to see. For network monitoring, I would rate it a nine out of 10 because you can collect all the information and slice and dice that information in whatever manner you feel necessary to consume that data. We've got an operations team that subscribes only to the alerts. So, we've got tier two and tier three people who are looking at reports, and they slice and dice those reports however they like.

Its collection abilities across multiple vendors' equipment are really good. If we don't have an SNMP OID for a particular vendor, the only thing that the architects at my company need to do is to supply us the SNMP OIDs and/or MIBs. We send these to SevOne, and they certify it. We can then install it in the SevOne system, and it'll start monitoring that equipment. Its collection abilities are important because we've got multiple vendors in the network, and each specialty, such as a firewall or a router, has different collection needs. We're able to meet these specific collection needs based on the device types.

For our operations, the dashboard is very important because that's how our customers are making day-to-day and long-term strategic decisions, for six months to a year, about their network. We're not using any reports for capacity planning as such, but this is an idea that is going to be put in place shortly.

It provides continuous analytics of the network, which helps our customers in making smarter decisions and ensuring that things are up and running.

In terms of the integration of network performance management data with our ITSM tool, we don't have a direct integration with ServiceNow. We have integrated SevOne with Netcool, and Netcool is integrated with ServiceNow. It is pretty easy to integrate. We've got people on our team who are responsible for Netcool, and if we want to define a new policy or alert, we show them what alert we're sending over, and they integrate it in a matter of a couple of hours.

Its ability to monitor practically any type of network device via SNMP is most valuable. This is the main functionality that we're using. If a network device exposes a metric, such as interface utilization, SevOne will monitor it for us.

The reporting is very good in SevOne. We have static thresholds that are defined by our architects. They give these static thresholds to us, and we implement the alerting policies based on those static thresholds. We also have the capability of doing base-lining or deviation from normal or mean, but we haven't implemented that in our network. 

The out-of-the-box reports are of quality, and they would get you up to speed faster than having to build custom reports. I wasn't here when the reports were created, so I haven't, as such, used the out-of-the-box reports.

We are able to use SevOne's analytics, reports, and workflows in a single dashboard. Its dashboard is very easy to use and put together. It is also really easy to understand. If I had to give it a grade, I would give it an eight out of 10.

In terms of having a complete view of our network performance, I would rate it a nine out of 10. The reason for not giving it a 10 is that there is no packet capture associated with SevOne, but we do have other tools in place to do that.

In terms of stability, because of our move to VMs from physical appliances, some things have become a little unstable. It doesn't seem to be a SevOne issue, but we had to have a lot of calls with their technical support to figure out what's going on with it, but overall, it is pretty solid. 

I have been using this solution for one year and two or three months.

Overall, it is pretty solid. We've made some changes to the SevOne infrastructure, and we moved to VMs from physical appliances. Because of this transition, some things have become a little unstable, but we're working on these issues. It doesn't seem to be a SevOne issue, but because of the change of infrastructure of SevOne, we have had to have a lot of calls with their technical support to figure out what's going on with it. 

It is extremely scalable. We're managing almost 8,000 devices, and if we need to add 8,000 more devices, we just need to add a commensurate number of peers to handle that load. It is horizontally scalable, which is nice.

They're readily available, and they work with us in a very friendly way. They are very willing to help us. Some support desks, especially in performance monitoring, push you to solve your own problem, whereas SevOne's support is the exact opposite. Everyone I've worked with has been helpful. I would give them an A. 

I think they used HP OpenView. I have no idea about the reasons for switching.

I was not involved in its initial setup. For its maintenance, we've got two people in the US and two people in the Philippines who help us. They do network monitoring. The two people in the Philippines work part-time on it because they also support other tools. So, we have three people in total for 8,000 devices.

I would advise evaluating it thoroughly to make sure it is right for your network, and it meets your administrative needs. This should be a major or key element of your decision process.

SevOne supports software-defined and streaming telemetry-based networks, but we are not using any of that. I've also not customized out-of-the-box reports. I've only created custom reports for various customer groups that are consuming the data.

I would rate SevOne Network Data Platform a nine out of 10.

We use the solution for performance and capacity management reports. The product gives us flow data that helps us determine the top users. We also use the solution for LAN, WAN, and Wi-Fi.

I like the tool’s scalability and real-time reports. Earlier, we struggled to give real-time reports to clients. I also like the tool’s deployment model where we can deploy it either on-premises or in-house. We don’t have to carry the data all over the globe. Also, I am impressed with the tool's flow reporting and Wi-Fi.

The tool needs improvement in non-Cisco SD-WAN.

I have been using the product for four years.

My company has 500 users for the solution.

We talk to the tool’s support on a daily basis or whenever we need their help. The product’s support is good.

The tool’s setup was easy. We were able to onboard two major customers within two months of the product’s deployment. The overall deployment took around three months to complete.

The tool is not expensive. We were able to negotiate with SevOne on pricing.

I would rate the solution a nine out of ten.

Sometimes we get requests that a customer needs CPU or disk or memory performance or utilization graphs. We add those servers or devices into the tool and then we can generate the graphs and provide them to the customer.

Customers also ask us to create alerts. The tool generates alerts for CPU utilization when it is close to, for example, 90 percent utilized.

It is deployed directly on servers as well as on virtual machines.

One of the benefits is its ability to transform raw network performance data into actionable insights. That's one of the keys for us. When something goes above a threshold, we can see it in the alerts and take action. Likewise, we can see graphs and reports and we can judge what to do before something goes wrong.

Some of the teams that are using our graphs from SevOne, and the capacity team that uses the data it generates, are able to make decisions before something bad happens.

We use SevOne to monitor a multi-vendor network. We have a lot of different kinds of devices in our scenario. We have Cisco switches and network devices from various vendors. The alerting and reports that we can generate help us see if something is not the way it should be.

• Reports
  
• Alerting

These are the most valuable features for us because the customers in our company primarily want to see performance and usage graphs, and they are always concerned with the alerts.

Another useful feature is that SevOne gives you real-time insights into your network performance. It polls every five minutes. That is important for our customers because there are some network teams that are always monitoring their networks. There is an option for setting the polling frequency to less than five minutes. That means you can monitor your infrastructure faster and we do that for some of our devices.

And the data collection functionality, using SNMP protocol, is good. It's doing its job.

I'm not really sure if this was the software's fault or a server issue, but a couple of years back the disks were failing on our SevOne physical server every month and the server would go down. The secondary server took over from the primary until the disk issue was resolved. That was annoying.

I've been using IBM SevOne Network Performance Management since 2013.

It is resilient. If the server goes down, all the data and functionality is taken over by a secondary server. In our scenario, there has been no data loss.

You can add as many devices as you want, but I think you need to buy more licenses to add more devices. But scalability is not an issue. We have seven to eight clients and we monitor more than a thousand devices for each one.

There are no new clients in the pipeline, but if another comes along, we will definitely recommend SevOne to them.

The SevOne support team is very good. Whenever you have a strange issue or a big issue, something you have never seen before, when you reach out to them they are always available. They are very fast and always help us.

Positive

I deployed SevOne on a virtual machine a couple of years back. The deployment was easy and straightforward. The installation wizard helps, giving you all the details of what is happening. There was no confusion. And it was fast as well. It took roughly two hours.

Someone from our deployment team helped me. He told me to just "apply this, do this, do that," apart from what the wizard showed me. I believe he was in touch with the SevOne guys.

I can't say anything specific about the investment in the solution because I'm not given that data by my company. But our clients are still using this solution after many years with our company. That is a good indicator that they must be getting a good return.

There are other tools that we have used, like eHealth from CA. It also gives you graphs but we don't like that tool. We like SevOne. It is older than SevOne. We are using it for some clients that have had it from the beginning, so we cannot remove it. But eHealth has bugs. With SevOne, I don't have any complaints.

Definitely go for it. The interface is user-friendly, and it provides so many reports and alerts. It gives you a good, total package. And the support team is also very cooperative.

I can't think of very much that the solution lacks. Everything looks okay to me.

My use case at the initial startup was very simple. I had a carrier, which was a backbone globally implemented, and I needed a monitoring solution. The type of solution I needed had to capture SNMP traps, poll my equipment, perform traffic analysis, deal with historical data, and things like that. This requirement has remained constant through the entire seven years of implementation with them.

At the end of the month, we're ending our relationship with this vendor for a variety of different reasons. Among the problems is the pricing model that they have, although a lot of it has to do with the fact that their virtualization solution isn't compatible with our Kubernetes environment.

SevOne has enabled us to integrate network performance management data across ITSM and our business decision-making tools, predominantly through the ServiceNow platform. We also did a Salesforce implementation where SevOne leveraged Salesforce to determine if a circuit was production versus non-production. Essentially, this distinction implies whether we should care about it, or not.

The integration with Salesforce was pretty easy, where most of the work was on the Salesforce side. It was probably one of the simpler integrations that we did for the platform.

The comprehensiveness of SevOne in terms of collecting network performance and flow data, when we started using this in 2013, was very limited. It was developed predominantly for a Cisco network and I'm a hundred percent Juniper. As such, it required a lot of work to get the platform to not only understand it but to speak in terms of Juniper MIB files, and even the nomenclature. For a Cisco network, it would have been a situation where you opened the box, plugged it in, and walked away. With Juniper, it was very much not that.

At this point, our collection capabilities are limited to just Juniper equipment. This is restricted by the tool that we have, which only covers Juniper networks.

With respect to streaming telemetry, we do not have it implemented. We were working with them to try and understand what they could do in this regard, but I do not believe that they supported streaming telemetry at the time.

The most valuable feature as of late has been the API integration with ServiceNow. Honestly, the biggest bang for the buck I've got out of SevOne has been this development. The bi-directional integration with ServiceNow has saved me a lot of money in man-hours, over the course of the last few years.

I don't have an exact figure for how much money I have saved, but I can say that it's hundreds of thousands of dollars. What it comes down to is when you're able to automate the console work with the ticketing system, you're saving people from copying and pasting, and other such menial tasks. For example, you are able to auto-populate tickets, update tickets, change the status of tickets, and also do verification to see if something is valid. You can make determinations such as whether there is a ticket currently open or whether there was a ticket previously open. Automating things like that, so a human no longer has to do them, can save hours a day per human per shift.

The out of the box reports and workflows are very sufficient for helping to understand what's normal and abnormal in the network. Out of the box, the reports were certainly there and even though it didn't necessarily understand Juniper, the minute we turned it on, we had a bunch of data. In fact, there was a lot of data that we had never previously seen before on the backbone, made available to us just by virtue of turning it on. It just needed to be cleaned up and polished.

We were aware of the reporting when we decided to implement SevOne, as we had done a lot of pre-sales work with them to make sure we knew what to expect out of the box. Even if we needed to do a lot of customization, it was certainly expected, and that's what we saw. It was important to us because we needed to immediately show some sort of value with all of the work that we'd invested over the course of the implementation. I needed to show almost a day-one value, and that certainly did help.

With respect to customization, the reports themselves didn't take too much effort. We have had a resident SevOne engineer help manage the platform and tend to those apps throughout the entire implementation of SevOne. From my standpoint, it was simply a case of asking the resident engineer for what I needed or what I expected, and whether it was a function of hours or days. Shortly after, I would have exactly what I needed.

An example of how we have customized reporting is the top talking report. It is important because we have a lot of customers that are very bandwidth-intensive. This report is for aggregate bandwidth and it is from a trap-generation standpoint.

I also have a performance metric where we monitor a specific group of circuits that are notorious for having capacity issues with customers. Essentially, it is a top talker traffic graph where I get the top ten circuits for the past 24 hours, and it's a live graph. I get it as a report, but I can also watch it in real-time.

SevOne provides continuous analytics of our network and it's important because if you're in a network where you're polling every three minutes or every five minutes, then you could be missing important events. There's a lot of stuff happening and it can be very damaging in a matter of seconds. If you're not polling or collecting data to absorb that frequency or that duration, then you're not doing anything. You're completely overlooking the important stuff. Being able to see in some form or another, not always in the graph, but being able to see that real-time activity and have it called out to a human is exceptionally important. Again, it doesn't need to be a graph, but that's one of the things we leverage SevOne for.

With respect to giving us a complete view of our network performance, it's been very good. I don't know how many times a week I have a STEM vice president come to me and ask me what's going on with the backbone or how the backbone is performing with a certain world event or corporate event. Whatever it may be, I can get a very good visual summary, very quickly, just by virtue of logging in. It's just a matter of making sure that you have the right graph. You have to tell SevOne what you need and have it presented to you in the right way. Otherwise, it doesn't know. Once you accomplish that, it's immediate.

SevOne has enabled us to detect network performance issues faster, and before they impact end-users. It is very good at capturing those events, documenting them, opening a ticket, and letting a human know about them. There is a definite ability of proactiveness with the tool.

If I consider where we were in 2013, it could take several hours or days to detect events in some cases. I have examples of catastrophic events happening that we never even knew about, that SevOne is able to capture. I estimate that we are 60% faster on average at capturing and actioning events, hopefully proactively.

Their virtualization solution is not compatible with our Kubernetes environment, which is one of the reasons we are ending our relationship with them. I didn't spend a lot of time evaluating with them why it was the case. It was simply not a roadmap item for them, so it was a pretty quick conversation.

We have been using SevOne for approximately seven years.

This product is very scalable, especially if it's just a matter of growing the network. You add more devices, make sure your licensing is in check, and the system ingests it as that equipment is green-lighted.

If you're changing technology, adding layers upon which you want them to monitor, it is still scalable, although it takes a little bit more work.

We have approximately two dozen users in the organization.

Their technical support is very competent. We have had an immediate reaction to our issues, even without the resident engineer involved. Their technical support is 24/7. That said, I've actually had very minimal interaction with them, aside from some hand-holding during software upgrades. Other than that, the platform has been rock solid.

Prior to using SevOne, we were using an internal homegrown solution.

After we got done building it, it largely sat idle until we started onboarding customers. As customers grew, a need for a focused operations group, tooling, processes, and procedures arose. That's where SevOne came in. We needed a legit platform to monitor the backbone rather than use existing processes and procedures that just didn't work or didn't apply.

Essentially, with the growth of the backbone and the responsibility of it, we realized that we needed an enterprise-grade solution.

The initial setup was pretty straightforward. We knew that the biggest hurdle we had to overcome was the Juniper compatibility, so that's where we focused the resources in the planning.

The means of actually getting it installed, upgrading the software, and then actually discovering the network worked as expected. It crawled, it discovered, and it did everything we needed it to. It just needed to be tuned for a 100% Juniper network.

Of course, the Juniper tuning took many hours of post-sales engineering support as well as a resident engineer. It took a lot of work on the SevOne side to actually get it to that point.

In total, the deployment took approximately three months.

I and a colleague were responsible for deployment.

Maintenance requires one FTE.

In terms of ROI, I don't have a whole lot in terms of metrics. However, I would say that with DI, someone has definitely started to come around from a visualization standpoint. Not only do you get an alert with an indicative color like red, orange, or yellow, but it is well represented for different stakeholders. It is not only useful for the engineer sitting at the desk but also for the tier-three that supports that engineer, all the way up to the vice president, who just wants to know how things are going.

They've come a long way in developing that. Back in the day, all people wanted was something that told them the status; red is bad, green is good, yellow means that you should look into it. That was all the information that they had. These days, people want predictive analysis and they want to be able to trend failure. They want to be able to dig into the numbers a little bit more and graphically represent that. To this end, DI is actually something that they're doing to chase that down and fill that void.

Historically, that hadn't been the case. I think DI came out approximately four years ago, and I think that's something that they're really doing to try and add value to the platform.

The pricing has not evolved with the market, which is one of the reasons we are moving to a new product.

When we implemented SevOne, we had been evaluating other options for a couple of years for varying needs, although not necessarily the backbone. During that process, we had noted that SevOne would be the most accommodating and capable for our needs.

At the time, it just wasn't possible for us to implement it.

SevOne is capable of bringing together its analytics reports and workflows in a single dashboard, although I don't actively use that specific dashboard. The stuff that I use with SevOne is very specific to a need at the moment and as such, I don't require the use of a collapsed view. In my world, it's hard to summarize everything in one place. Everything is going to be compartmentalized, so I have multiple dashboards with different data. It isn't that I don't want to use a single pane of glass but it just doesn't serve any purpose for what I need on a daily basis.

Overall, this is a good product and we had a really good relationship with the vendor. When it all started, I had a pretty basic need that I was unable to get any support internally for. We had spoken with them before, and at that initial time, I had some internal obstructions to bringing them onboard. The problems were not financially related and over time, as usual, things changed and the obstructions were gone. Once that happened, I was given the opportunity and the power to develop my own tooling suite for my team, and SevOne was a pretty easy discussion at that point in time.

The relationship continued to be a really good one up until a couple of years ago, when we were growing and of course, they wanted in on that, but their pricing was not adapting to what we were seeing in the market. They were still doing pricing from 2013 when we bought in. Naturally, anytime I expand tool usage, it works in my best interest to make sure that what I'm using is still the best implementation for not only the cost but also, the scalability at the time.

The biggest lesson that I have learned from using SevOne is that leveraging your platforms to do more work in place of a human, isn't always a bad thing. A lot of people think that you're just trying to replace humans with automation and software. What it really boils down to is that you're enabling those humans to do something else that is more important. It's not a function of eliminating jobs. It's letting the humans work on more important, complex items, and let the software and the automation do what they can to contribute to that equation.

It's not that it's necessarily been a challenge or an obstacle for me, but it is important to consider it when explaining the process. When you explain to someone that we're changing this process because SevOne can now do a certain aspect of it, with human involvement starting somewhere further down the line, you have to be able to sell that as an improvement to the process. Ultimately, it's allowing that human to focus on other things that have previously been neglected.

This problem of automating a task that is historically done by a human has been a lesson that I've learned with SevOne. The reality is that you have to let automation do what it can, and let humans do the more important engineering work. Getting away from that stigma and letting the software do its job and really focusing on releasing that, allowing the humans to do the more technical and engineering-level work, is really an act in cost-savings and from a Human Resourcing standpoint, you're getting more bang for your buck out of it. You don't want to pay people a lot of money an hour to sit there and say that red is bad and green is good. If you can get away from that, you're going to be more efficient.

I would rate this solution a seven out of ten.

• Flexible architecture: You can extend the system and its capacity by attaching another cluster pair.
• Very intuitive management interface: Adding and discovering new devices is a very simple process.
• Very useful and flexible end-user GUI interface: Reports or statistics can be prepared by a person who has no knowledge of performance monitoring.
• Automatic reporting: You can very quickly prepare a report to be periodically sent to recipients.
• Very fast reporting engine: Even very complex reports are generated in seconds.
• Many predefined Top-N reports are available out-of-the-box.
• Grouping capability: Each device can be assigned to many groups, which means you can report any interesting network factors according to the multiple group allocation.
• Baseline: The functionality that allow us to monitor a particular factor (like throughput or CPU load) based on some historical data (the value of the factor at similar period of day should be more or less the same)
• Virtualization of network elements: Many physical interfaces that exist on different physical devices can be aggregated as a single logical device with many logical interfaces. This is very useful functionality for network operators.
• Trend analysis: The system predicts the value of the traffic in the future based on existing behavior.

We provide customer internet access services and the 95th percentile is our target. Every month, we prepare a detailed report per customer that shows the current percentile value (does it exceed 95 or not), and we have to prepare detailed traffic reports that show the real traffic graph in the last month.

All of this was done manually. With SevOne, this process is fully automated and the reports can be sent directly to business customers (after a simple verification performed by another colleague).

Our version is quite old. In version 5.3.3.0, we see a lot of room for possible improvement. However, from SevOne support, we received confirmation that most of those expectations are met in version 5.4.x or higher.

Therefore, we have to think about upgrading to the later version as soon as possible.

SevOne should work with the graphs legend functionality. Now, you are able to put a part of the graph description as a customer description, but most of the original description stays unchanged. This means that sometimes the legend under the graph is unreadable (indicator names are sometimes not humanly readable). It would be nice to have a solution similar to the one in Cacti, where you can replace (or rather overwrite) an existing description with your own string.

It has now been almost one and a half years.

Currently, we are using version 5.3.3.0, but we are now just upgrading the system to version 5.3.10 due to several minor issues.

Not, over the last one and a half years.

Yes, but this was caused by special and very uncommon expectations from our side.

Together with SevOne, we implemented the solution which allows us to automatically add any new network device that is added to our external (independent from SevOne) database.

We use the SevOne API to add those devices and interfaces to SevOne. In case of devices with a huge number of interfaces, more than 100, SevOne was not able to load them into its own database.

SevOne recommended an upgrade to a later version to resolve this limitation.

This is the part I am happiest about. Their response is great.

Of course, I sometimes have to wait a week or two, but mostly that is because of the nature of the problem and its complexity.

Most problems are resolved within two to three days.

In our case, the SevOne platform was implemented by a third-party integrator. So, at the beginning, our contact with SevOne was very limited.

Now, for simple or medium issues, we contact SevOne directly through the SevOne support webpage because it speeds up the problem solution time. However, for more complex issues, we still contact the third-party integrator.

We were using Cacti and Zabbix (both open source solutions). We decided to switch mainly because there were some business expectations to have a platform that would prepare reports we can show to business customers.

On the other hand, we would like to have a tool ready to prepare reports on demand and by the non-technical staff.

In a standard solution, the instalation is very simple. In our case, we decided to integrate SevOne with an external database (an external application). All network devices devoted to that application should be automatically inserted into SevOne database.

The integration interface was the part of the solution that was performed by a third-party integrator in cooperation with SevOne.

From the operator’s point of view, it is quite painful to have to remember that every device costs us some cash if added to SevOne (CAPEX and, later, yearly OPEX).

Prices per license are not huge, but they exist.

It is very visible when we connect a big number of network devices (due to some new company acquisition). At once, we have to connect 100s of network elements, and it is hard to find extra money for that.

On the other hand, the existing model is very flexible from a financial point of view (pay-as-you-grow).

Yes, but there was a tender, and I am not authorized to provide such information.

I can only say that there were a few big players in the area of performance monitoring system vendors.

• The grouping capability is very simple and a very important issue in terms of system reporting capability. You should do your homework carefully so you will have a flexible reporting tool in the future.
• Enabling baseline functionality: We decided not to enable it at the beginning and very quickly decided to change our minds. It is a very useful mechanism for data comparison (today’s traffic to the week’s traffic, to weeks before at the same time, and so on).
• Report preparation: It depends on the agreement, but SevOne is ready to prepare some predefined report at initial integration. Let them do this to save you time, but it requires some time to think about your expectations.
• In SevOne, you pay mainly per object. Do not enable all object pooling by default. In a case with 10,000 devices, if you decide not to pool ICMP (not to ping devices to check availability), you can save 10,000 objects, and save real money. (We did so and we do not regret that decision, but it depends on the particular expectations of the company implementing the product.)
• The same as disabling ICMP pooling, you can decide to disable memory and CPU monitoring (if it is not necessary). Money that again stays in your pocket.

The platform is as flexible as an open source system can be. It has a very useful end user GUI interface, which means working with the system is very easy and intuitive.

We use it to look into performances in our system and different dedicated environments. We try to draw graphs that represent what we have in terms of time response, bandwidth, and input and output across all our data centers to see if the load balance is correct. We have some traffic flows that say if the traffic is dropping and being taken by another data center. There are a lot of reports all around about different points of our architecture.

It is deployed as a physical appliance on-premises. So, we have the SevOne appliance in our data center.

Based on some metrics, we have policies that will raise issues and tickets internally, in case thresholds are being crossed. 

We have benefited mainly from the use of the dashboard interface. It makes the network visually interesting for people who are not in the network. A lot of people are not network techies who understand streams in the network. Based on location, we have streams coming in and out. They can see visually when there is some problem. They don't need to understand all the network technology behind it to be able to understand if everything is working well or if there is a problem. 

The visualization doesn't remain in the network. It is available for everyone in production management, etc.

We don't use all the features so far because we are just using the product that was there without supervision for one year. So, with a couple of colleagues, we are doing a lot of work to put everything back on track. We know that SevOne can do a lot, in terms of hitting some baselines, especially with their new visualization software, SevOne Data Insight (DI). This gives a new perspective in terms of dashboarding. It is something more dynamic and flexible than the previous version.  

There are different types of reports based on the metrics that we want to monitor and check.

The most valuable feature is that we can draw reports with the historic value of data. So, we can see if there is a trend in the past or if something has been changing over the past couple of weeks. When we have an incident, we need to go back to the occurrence of this incident some time in the future. We can see if it is something that happens regularly or not.

Data coming from our system to SevOne needs to be comprehensive. This is the way SevOne works around data flows. So far, it has been good.

They support software-defined networks as part of another module. They have their main module, which is called NMS, then they have other side software which complements the main architecture. They also support software-defined networks, like Cisco ACI. 

We made assessments internally about the system, i.e., about what could be done better. There are a lot of points. 

With the administrative management of the appliance, if some object appears from SevOne because something changed in the network or whatever, then as an administrator you will not be aware. If you are using this object in a report, this object will disappear from the report and you will not be aware of it. So, if you have 1,000 reports, you cannot always check these reports everyday to see if objects are missing or information has disappeared. We don't have any information on alerts, saying that something is happening there and maybe we need to take action. If an object was replaced by another one, or if a link was replaced by another one, then the graph needs to be changed because it doesn't exist in the graph anymore. However, we don't have this information.

This is also the same in terms of the internal architecture that we put in place inside the system. We can tag our network device based on the firmware, some rules, or a manufacturer/vendor. But, it is not always clear when we add a bunch of devices that they will mark these categories. For example, we need to make sure that one device is from a specific vendor. We have to dig deep in each device to make sure they are really attached to the correct vendor. If they're not attached to the correct vendor, then information will not be pulled in the same way, and we might be losing information. So, small tweaks need to be made to the internal management, making it easier for me.

SDN networking is going from legacy towards object-oriented, so we don't have a choice. It is something that we are using and need. Unfortunately, our IT lead is not really into the SDN solution of ACI, as apparently it is missing some features in terms of deep analytics and monitoring of ACI hardware infrastructure as-is. In the future, we may use SDN infrastructure because I know they are releasing features every month.

We have had it in service for five years. I was not there when it was installed at the beginning. I have been handling it for a year and a halfs.

It is really stable. We don't have any issues, except when we do our disaster scenario a couple of times a year. We have to physically cut the connection between our sites for redundancy and compliance. Then, we can sometimes see that we need to ask the help of support to put back the database in a correct state because the replication is not always clicking well. But, that is most likely because of the technology used behind it. MySQL is always a bit sensitive in this kind of scenario. So, I am not sure it is really the SevOne application that has an issue. It is more the technology that lies behind it which can be a bit faulty in these kinds of scenarios. Besides the replication, it works really nice. 

Without disaster scenarios, we are reaching 100% of availability easily.

We just scaled it up a month ago because we were reaching some limits. It is quite scalable. It took a day to scale it up with the new appliances that were delivered. It is quite easy to scale.

We are monitoring 886 devices. That is infrastructure monitoring. There are more devices that we still want to monitor. Our plan is to reach 1,300.

We use the technical support quite often. It is good, though it depends on the location that is doing the support. There are several locations that do support, whether it's India, Poland, or the US. Depending on that, the support can vary as well as the response time, because we're not always in the same time zones as their support. Sometimes, it can take a long time to receive an answer if we have been redirected to the wrong support.

Since it is something new, we are in contact with the vendor. We can give them some updates about what issues we are facing and if we are happy with the product so far. 

We are using another tool called EMC Smarts, but it is more for event monitoring. So, there isn't any drawing of dashboards. It is more only thresholds for SNMP Traps that can be received. Based on that, it distributes tickets. So, we cannot see the history nor draw the status of a link. 

A long time ago, I used Nagios, which is kind of the precursor to monitoring. However, I cannot compare it to SevOne because I used it about 10 years ago. 

You have to try it. The first step may be a bit harsh because of the layout. It's an older type of layout, but it works. Once you find your way with it, it is quite easy and straightforward. But, at first, it's not always easy to handle. But, it is a good tool that works nicely. So far, we don't regret it.

We need to master the system to be able to fully grasp what can be done. It is not for every end user who wants to go into the system and start monitoring whatever they want. They need to be able to grasp some content before that. Without training or previous knowledge, it is not always easy to grasp every concept behind the solution. It is a monitoring system so it's not a system where you click, then drag and drop.

I know they support telemetry. But the device on the end user side needs to be able to send telemetry information, which is not something we can do yet. We don't have enough devices doing telemetry to really use that feature to its greatest potential. So, this is something we have on our roadmap, and probably we will dig into it in a couple of months or years when our infrastructure is evolving in that direction.

Telemetry is something that we would like to invest some more time with, because it is different from having just simple SNMP polling, which is heavy on the system. It puts on a lot of overload based on the frequency of polling. By default, it takes five minutes. We want to have something with a frequency smaller than five minutes and maybe pause every 10 seconds, and SNMP can do it, but it puts a lot of overhead on the system. With telemetry, the big advantage of telemetry is a constant stream of information. There is no overhead. We just have a constant flatline of internal usage. We don't have huge peaks. We have fast information, close to real-time. We should have closer to real time monitoring in the future, instead of just being passive and waiting. There is still a ways to go for telemetry, but most infrastructure is capable of doing it.

Nothing is perfect. I would rate it a solid seven (out of 10) because a lot of points could be improved. I have a long list of small tweaks and customization that can be brought to the system. We give some of them to the customer support, but we are not their only clients. So, they go through prioritizing all their processes. Sometimes, our propositions are refused. It is a good system, but there is still some room for big improvements.