Looking at POC for CA Nimsoft wanted to see if anyone has moved from CA to Solarwinds and Why?
Nimsoft is a great product that brings a lot of unique features to the table. It is a different beast than Solarwinds, and there are both benefits and disadvantages to their approach to monitoring. I am a Solarwinds admin and a Windows admin, but recently my company was considering standing up a Nimsoft environment so my team scrambled to work with CA and learn as much about the product as possible. Here are some things we learned..
Nimsoft is a multi-tenant monitoring solution by design, Solarwinds is not. Solarwinds can claim multi-tenancy, but if you look hard at the structure of the NPM database and the design of the web console; it wasn't built in a multi-tenant fashion. Nimsoft was built to maintain a separation of data based on company.
Nimsoft is almost infinitely scale-able because of its message bus, Solarwinds has limits. The Nimsoft agents report monitoring data using a store-and-forward message bus. They function in a tree-like structure to pass messages up the chain until the data reaches the main database. This means that not all data is sent directly to the database, but the load is spread out among all the clients. Messages work their way up to the database gradually and it happens to allow for Nimsoft to take on a way larger scale than Solarwinds. There is built in fault tolerance too, each Nimsoft agent is configured with a primary and secondary recipient for its messages. If a node goes down, the agents will simply reroute messages along the bus.
Nimsoft is agent-based, Solarwinds is agentless (although technically WPM employs an agent, and creating a SAM agent for Windows monitoring is on the roadmap). Agentless monitoring has its merits, but it depends on the existing monitoring protocols already out in the wild. It does allow for us to do cool stuff like Network Sonar discovery of nodes, and pretty much gives Solarwinds the ability to start monitoring an environment for ping up/down right out of the box without any client configuration at all! Agents are more invasive to client machines, but they allow for a more secure and unrestricted way to report data. Microsoft has deprecated SNMP, which is a major concern going forward. Polling via WMI is a nightmare, as it demands too many open ports and is not as light-weight as SNMP. It seems like Microsoft is pushing things towards WinRM, but it becomes challenging to securely monitor a server via WinRM that isn't on the same domain. All of these concerns are addressed by having an agent.
Nimsoft is overpriced and their pricing model is criminal, Solarwinds is more reasonable. If you are looking for a good reason not to switch to Nimsoft, just look at how much it will cost. The worst thing we found about their pricing model is that it scales based on the amount of nodes you are monitoring, and your Nimsoft installation calls home to report how many. Nimsoft doesn't just monitor everything out of the box either. They call the engine that drives the actions of its agent a 'robot' and each action requires a specific 'probe'. There are probes for anything you would want or need, such as basic ping monitoring, tracking services, and even storing and forwarding monitoring messages on the Nimsoft message bus. They bundle the probes into 'probe packs' and make you buy more packs if you need to monitor different things. So the more diverse your environment gets, the harder they hit your wallet. Solarwinds seems to price based on a more broad stroke. Between SAM and NPM, there isn't much you can't monitor, and the only increases in cost come from having to buy additional pollers.
I hope some of this helps, but its worth noting that I haven't actually used Nimsoft outside of our testing environment. I prefer Solarwinds to Nagios or Nimsoft. Solarwinds empowers its admins with tools to create solutions, monitor anything, and even interact with the database or website. Any problems we have had with Solarwinds, we were able to overcome by using custom properties or implementing our own bolt-on custom solutions.
Disclaimer: I've attempted to view, use, and RFP Nimsoft in two previous jobs, but for various reasons never actually saw the tool in action.
From what I understand (a neighbor of mine is an SE for CA and works with their monitoring products) Nimsoft is agent-based solution. Similar to Tivoli, what you are mostly getting are the agents, the 3-tier architecture that allows agents to communicate back to the central data store, and the software infrastructure that managers those agents. From there, you code up the monitors, alerts, or response actions you want. I don't know if they have any (or how many) pre-configured options for monitoring they have.
What I do know is that it is fairly complex to deploy and maintain, and that it is significantly more costly than other options (such as the company that owns thwack.com, just as an example!)
Speaking as someone who has worked in companies where monitoring was done with tools from BMC, IBM, and Microsoft I can tell you that the cost to purchase, effort to maintain, and amount of staff required have always been greater (double to triple the number) for agent-based solutions versus agentless.
Leon Adato while I agree that having an Agent certainly adds some overhead, in my experience with the agent based solutions it's less the agent that causes the additional overhead and more just the way those specific products are designed in general in that they are huge and incredibly (and unnecessarily) complex products.
With that being said, I would agree that Nimsoft is certainly a more expensive product than a comparable product that rhymes with "Orion". Back when we reviewed monitoring products for a new enterprise solution we looked at Nimsoft and the pricing model alone was enough to keep us from looking any further.
Good point.
Agents don't kill monitoring teams. Monolithic companies with bloated pricing structures and 90's era toolsets that have never been updated do.
Well, that and giant flying hyper-intelligent sharks with frikkin' laser beams strapped to their heads.
sneakernet Great explanation! This covers everything I remember about the details we were provided by Nimsoft and more.
One of the reasons we had originally looked at Nimsoft was due to the native multi-tenant support as I agree that Orion is not a true multi-tenant solution; however, we do use it as such. This has been a continued frustration with Orion and I have continued to push SolarWinds to work on this as it seems it wouldn't take much for them to get there.
Again, thanks for the great explanation!
I vote that sneakernet's answer is correct. Go ahead and tag it so he gets his points! Great work!
Thank you all for your feedback. This is depressing that my company is planning on other tools since I have grown quite fond of SolarWinds and the Thwack user community. The first thing I looked at with CA was the user forum to see if there was any comparison so far there is not. The biggest driving factor I am facing is that the company has selected CA Service Desk as our new helpdesk system. CA Nimsoft is promising on full level integration stating that the following are able to be put in place. Help desk ticket open/close/update tie in to CA Nimsoft alerts as well as CMDB and Change Control maintenance options. We are phasing out Service Center (HP) and we already have the option of opening (not close/update) with SolarWinds. We are not using a common CMDB or Change Control with the HP Service Center which is in place. This is the main selling point as well as the wholesale one product one vendor approach that I am looking at having to overcome to keep SolarWinds. Your feedback and individual experiences are very insightful and helpful so for that I thank you.
I thought I would chime in on this as I work for a large MSP and we just recently went through the comparison of these two. A little background on myself I currently manage all tools used by our Managed Services group and have been in the infrastructure monitoring game for 6 years. Before that I ran a SOC for a large retail company, same approach just less tinfoil in infrastructure monitoring. J It is important to also say I am loyal to my customers first and not vendors. This is a statement I say to every vendor “If another tool comes around that is better for my customers, I will not hesitate to rip you out”. As harsh as that may sound to the sales guy sitting on the other end of the table it sets a clear understanding of where my loyalty resides. Currently we use CA Nimsoft as our main monitoring tool but we do use supplemental tools from other vendors SolarWinds being one of them. We use Kiwi Cattools, Kiwi syslog and are investigating Storage Manager for deep dive storage reporting.
One of my biggest complaints with SolarWinds was the lack of multi-tenant capabilities and the lack of caring from SolarWinds. They almost have a snobbish attitude about it “If you don’t like it to bad, we will find someone else to buy it”. SolarWinds is so close and it’s frustrating that they could do it if they wanted. They would have a really compelling offer if they did. CA Nimsoft on the other hand is 100% designed for multi-tenant environments, which is crucial for a MSP or large Enterprise with multiple divisions.
One thing sneakernet hit on that I wanted to elaborate on. Is the HUB and SPOKE design of the infrastructure, this is a HUGE advantage. Nimsoft has two parts to this design a “HUB” and a “Robot”. Both HUBs and Robots collect information but robots are dumb and push that information to their HUB. There is one main HUB for the entire environment but you can have unlimited child HUBs. These child HUBs can act independently from the main HUB. For us we use this buy putting a HUB in each customer’s datacenter. This means that datacenter is fully monitored (alerting included) regardless of the state of our main HUB. This means I do not have to sweat bullets when our environment needs maintenance done to it. These child HUBs still report back in to the main hub so it is centrally managed from one console on our end.
The point that “Nimsoft = Agent and SolarWinds = Agentless” is incorrect. Nimsoft can be Agent or agentless which is crucial as both have their pros and cons. From a deployment standpoint both are easy to deploy with Nimsoft. You put in the networks or IPs to scan, credentials to use and it builds an inventory and reports back what it was able to log in to (Windows, SSH and SNMP devices). With a few clicks you can tell it to deploy agents or add it to agentless monitoring.
As far as ease of use this is my next big complaint with SolarWinds as I cannot stand the portal. SolarWinds buys a bunch of really cool products and “integrates” them together. The issue is you can tell when using the portal that you are moving from one product to the other. This is not always an issue but more an annoyance. It does worry me on where you might run into limitation with this “Jammed together” approach. With CA owning Nimsoft they are slowly adding CA tools to Nimsoft and it will be interesting to see how this goes. When they introduced NetFlow it had the same issues but they have now made it more integrated.
I found it interesting reading sneakernet’s thoughts on the licensing model as I always felt like I was getting nickeled and dimed with SolarWinds. I am not sure if we have a broader scope of device types, but I felt like I needed to attend a class just to understand it. For us with Nimsoft we ask our customer “How many Servers, Application Servers (Exchange, AD, DBs), Network devices, Advance Network devices (Cisco Nexus) and Storage do you have?” That is five device types they need to count and we price out monitoring on it. Most leaders in IT should be able to get that number easily.
The last big advantage we have found with Nimsoft is monitoring is truly limitless. I know that sounds very sales guy like, but it is the truth. Nimsoft is built to be a sandbox and allows you to not be constrained by what a developer thought you wanted. This is a huge plus for us as we have customers that have legacy applications that have no output method, we simply built a synthetic transaction that logs in and scraps the screen for errors. There are backend commands that you can run against an ESX cluster to get the health of the environment that we are able to have Nimsoft execute and dump to a log which it scrapes and alerts on.
Don’t get me wrong SolarWinds is a great tool and I know a lot of people are happy with it. Both tools ar in the top of their class compared to others in the market. I think it does a great job for single environments when you only need to monitor a single part of the infrastructure stack. It has some really nice management tools which we have customers use in conjunction with Nimsoft. I understand my use case may not be the same as everyone as my use case is centralized monitoring for multi-tenant, large scale environments and full stack monitoring. I know not everyone has these same needs but with our managed monitoring service I have replaced many SolarWinds implementations and heard nothing but good things from those customers. I do hope that SolarWinds address some of these issue as it has so much potential.
Hope this is helpful!
Thank you for your addition to the comment thread. I have some of the same concerns with regards to where I stand with tools. I am an employee of my company not a cheerleader etc for a vendor. What I see in my environment is 1600 devices (300 cisco and remainder UNIX/AIX/LINUX/Windows) they are pretty simple with what they want Up/Down/CPU/Memory/Volume and the ever present when, why, and how things went down so reporting and alerting plays into that. With the change out to CA Service Desk we now have a integrated function unlike before that must be considered. I am really have learned a lot from this Thread and the up coming P.O.C. will help cement how things will work in our environment based on our test cases. Keep the comments coming!
I like what prettyflyforawifi has contributed. With the Solarwinds pricing model, they do tend to squeeze you. It is efficiently priced for my specific environment because we are using it at a way larger scale than it was ever designed for, and we have been able to make the most of their unlimited licenses. The downside to what we are doing is that we have had to create our own solutions to the lack of multitenancy.
Another great point was that the web interface and the integration of different acquired monitoring tools can be frustrating. The NPM web interface is great, and they keep improving it. Also, they finally came out with a web-based report writer which is sleek and powerful compared to the server-side tools it is replacing. It all breaks down, when I have to use SAM. SAM serves a purpose that Solarwinds really does need, and I am excited that an optional SAM agent is on the roadmap. However, they bought this thing and bolted it onto Solarwinds and it is painfully obvious / frustrating. The database is structured way differently between SAM and NPM, and the potential to work in bulk barely exists. We have been able to create some cool custom reporting tools for NPM, because the database structure is very straightforward and we can build a lot of our own SQL queries to pull from it. The SAM database is very differently structured, and trying to query it is not as easy. There is also a strange anomalous relationship between 'Applications' and 'Application Templates' that Solarwinds has never properly defined. We have had to adopt a specific naming convention that includes the node name and enforce a strict 1 template per application rule to create some order. So if we have to make adjustments to 20 SAM Templates, it requires our lowest ranking guy to go in and commit those changes one by one. This is even more frustrating because our lowest ranking guy is me.
I don't consider myself to be brand-loyal to any product either, but I like what I see with about 80% of Solarwinds. For basic up/down/cpu/memory/volume monitoring Solarwinds or Nimsoft would work great. Nagios is another strong contender too.
