What is most valuable?
- 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.
How has it helped my organization?
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).
What needs improvement?
Our version is quite old. In version 126.96.36.199, 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.
For how long have I used the solution?
It has now been almost one and a half years.
Currently, we are using version 188.8.131.52, but we are now just upgrading the system to version 5.3.10 due to several minor issues.
What do I think about the stability of the solution?
Not, over the last one and a half years.
What do I think about the scalability of the solution?
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.
How are customer service and technical support?
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.
Which solution did I use previously and why did I switch?
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.
How was the initial setup?
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.
What's my experience with pricing, setup cost, and licensing?
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).
Which other solutions did I evaluate?
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.
What other advice do I have?
- 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.