What is our primary use case?
We use this solution for monitoring both service provider networks as well as customer WAN / LAN / VOIP networks, including all kinds of SNMP-based edge devices.
The key solution to be solved is the correlation between all types of messages from all devices. We are receiving more than 1,000,000 SNMP messages a day across many types of domains.
Unifying the solution within our organization created a standard way of working and integration. We needed a centralized redundant solution for all of our domains, which requires an SNMP-based monitoring tool.
How has it helped my organization?
We were able to standardize the internal processes across all internal departments, resulting in almost an elimination of non-standard process flows through our organization.
We were also able to standardize by cross-learning from all teams on how we should configure the monitored elements, align the configuration of the monitoring application, and use the solution such that it created a common understanding of the best practices.
By implementing the automatic service incident, a very strong closed loop is created in the monitoring of the domains.
What is most valuable?
The topology correlation for layers one, two, three, and four significantly reduces the number of false positives (>98% ratio). By using Spectrum, we did not need to configure the network knowledge about redundancy. Rather, it learns the latest correlation between the monitored elements automatically.
By adding the alerting from the Broadcom performance management tools, a comprehensive list of alerts is created.
As a service provider, we added views for our customers to inspect their sub-domain in our whole network, which reinforces a high level of trust within the relationship to our customers.
What needs improvement?
This solution is missing the in-depth SDN correlation due to the late arrival of Spectrum VNA features and supported SDN networks. The various setups of the SDN networks are not always supported (yet) by the Spectrum VNA engine.
Integration with non-Broadcom AIOps solutions is a known area. We want to make our own choice for the AIOps solution and do not want to be forced to use the Broadcom OI solution by default.
For our size, a redundant based docker platform should be supported for large scale environments with more than 80,000 devices. This will have a very strong positive impact on the effort we need to spend in maintaining our large scale Spectrum solution and the related cost. The time to market for the new Spectrum solution will be reduced dramatically if this becomes available.
For how long have I used the solution?
We have been using CA Spectrum for more than 15 years.
What do I think about the stability of the solution?
Stability is good, especially for large scale location redundant solutions with more than 80,000 devices in a single cluster.
In general, the location-redundancy with automatic failover is a myth, except if you are using Spectrum. If you setup spectrum in a redundant setup across 2 location (80 km in between), you can have a automatic fail over time within 10 seconds. This simplifies the maintenance of your setup a lot as well: any component can be switched off for maintenance purposes. The redundant instance takes over within 10 seconds.
What do I think about the scalability of the solution?
Scalability is very good.
How are customer service and technical support?
The support depends on the complexity of the problem. Because our own knowledge is quite high and our environment is quite huge, problems can be very complex and time-consuming to solve.
Which solution did I use previously and why did I switch?
We used several solutions prior to this one. The reason we switched is that the combination was causing a problem and we needed to standardize.
How was the initial setup?
You do need a decent level of knowledge to roll out this huge scale solution.
Small, single location-based solutions for 10,000 or fewer devices are relatively simple.
In our experience, the OOTB settings are very usable.
What about the implementation team?
We did implement this using our own in-house Spectrum specialists, which do have knowledge that is on the same level as the first line support at Broadcom.
What was our ROI?
Our ROI is realized through better NPS and a smaller labour cost.
What's my experience with pricing, setup cost, and licensing?
In general, the license cost will be about 30% of the total TCO for this tool (hardware/support team, OS costs, and OS management). It turned out to be cheaper than other solutions, even those in the public domain.
Which other solutions did I evaluate?
We evaluated IBM Netcool, different public domain-based SNMP solutions, HP OpenView NNM, and SolarWinds.
What other advice do I have?
We hope that the speed of the improvement and product growth will soon again be at the same level as when this was a CA product.
Which deployment model are you using for this solution?