Although it is perfectly possible to have a single solution, I'm not convinced that it is a good idea. I understand that it sounds great to have just one platform, but you should also keep in mind that no solution lasts forever, and the more you rely on a single system or vendor, the more difficult it will be to replace the system in the future.
Actually, it will also often cause issues with upgrades of software or hardware as there is too many things that have to be coordinated. Many companies end up running old non-supported systems, because the upgrade process is too expensive. Monitoring solutions typically do not generate a lot of revenue, and therefore it is difficult for finance/top management to prioritize this when you compete with new services etc.
I have also experienced that vendors that had a very customer centric and positive culture changed completely after an acquisition, and suddenly the positive experience became a bottleneck that caused delays and problems for launching new services.
I'm not a fan of going all in on microservices (the orchestration can be a nightmare) but splitting this into a couple of more specialized monitoring systems that pass all alarms to a centralized ticketing system would probably be a good idea.
Hi Eurica,While I agree that having a single solution / vendor has its risks in terms of vendor lock-in, on the other hand, it also has its own advantages from a seamless integration perspective. We have worked with some decent Telecom companies on similar projects including Airtel, Reliance, Maxis, Celcom etc. and our integrated Everest InfraonOSS solution has provided a lot of workflow automation with different components working together seamlessly.
The good thing is that nowadays most tools support 3rd party REST-API integrations so in case in future you want to change any particular component, its much easier than the earlier days of XML / Corba based integrations.Some of the highlights include:
Kindly let me know in case you need any further details, I will be happy to share some case studies and elaborate on how we have deployed our OSS / UNMS solutions to cater to heterogeneous networks monitoring with a common aggregation layer for fault consolidation and analytics.
Hi Sachin,The suggestion we give to our customers is to list down the individual feature expectations of the multiple components / modules that are required and then have a separate section for the analytical points which overlap these individual components. For example, having a IT-to-Service mapping report which collects data from application / network / server etc. components and highlights the overall Service availability and relates its availability impact to the performance degradation of any particular component. Another example is cross-domain correlation for quicker RCA (root cause analysis).We also recommend assigning priorities to all the features that are listed down in the overall requirements document to help during the cost-benefit analysis and also for big projects it provides clarity on a phase-wise solution delivery approach.
Hope this helps.
Its very much possible to have one system to monitor all but it's would require a cultural change at the organisation; reason being that the infra is dissociated from business and ownership also defined accordingly.
People managing business have least ownership of infra and vice versa. Both business and infra KPIs need to be married and seen as a collective success/failures.
Technology can play a big role in it along with a holistic mindset. While considering any APM tool, one must define both Infra and business/functional KPIs and correlation KPIs, before finalising a suitable tool (there are multiple tools available but due deliberation should be done based on application context and non-functional impacts).
Organisations must invest in people to customise these APL tools, if needed, to make use of IoTs, machine learning and integrate feed into chosen APM tool.