Splunk for SIEM: how Enterprise Security works

How Splunk works as a SIEM: Enterprise Security, the CIM data model, detections and risk-based alerting, licensing, and what to check before you migrate.

9 min read
Splunk for SIEM: how Enterprise Security works

Key takeaways

  • Splunk works as a SIEM through Enterprise Security, a premium app on Splunk Enterprise or Splunk Cloud Platform: it runs on data you have already licensed and adds no ingest capacity of its own.
  • Data is normalised with the Common Information Model at search time; SPL detections (formerly correlation searches) create findings, and risk-based alerting groups weak signals by user or host.
  • Splunk quotes Enterprise Security per workload, on ingest, workload (SVC or vCPU) or activity-based pricing, so both data volume and search load can move the bill.
  • Reviews usually start from the ES lifecycle (7.3 support ended on 28 February 2026, and the support policy sets 30 October 2026 as the end date for 8.0), data growth, the operational load of self-managed indexers, or a renewal.
  • A move to Google Security Operations is planned around CIM-to-UDM mapping, rewriting detections in YARA-L, deciding where history lives, and a parallel run before cutover.

The short answer: using Splunk for SIEM means running Splunk Enterprise Security, a premium app that sits on Splunk Enterprise (self-managed) or Splunk Cloud Platform. Data is normalised with the Common Information Model, detections written in SPL turn events into findings, and risk-based alerting groups weak signals by user or host. Licensing follows ingest, compute or search activity. Splunk has been part of Cisco since March 2024.

This guide explains how each of those pieces works, when Splunk is a sensible choice, why some organisations review it at renewal or upgrade time, and what to check if you decide to move to another platform such as Google Security Operations.

What using Splunk as a SIEM actually means

Splunk's platform is used for IT operations and observability as well as for security. On its own, it indexes and searches machine data. The SIEM layer is Splunk Enterprise Security (ES).

  • The platform underneath. Splunk documents ES as a premium app used in conjunction with Splunk Enterprise or Splunk Cloud Platform. You need a platform licence first, and ES runs on the data you have already indexed.
  • Two editions. Splunk currently sells ES Essentials and ES Premier. Essentials covers detection, triage, threat intelligence, investigation and case management. Premier adds Splunk SOAR and user and entity behaviour analytics (UEBA). Both are available in the cloud and on-premises.
  • The owner. Cisco announced on 18 March 2024 that it had completed the acquisition of Splunk, for approximately USD 28 billion in equity value.

How Splunk Enterprise Security works

Data in: indexes, add-ons and the Common Information Model

Logs reach Splunk through forwarders and other inputs, and land in indexes. Each source has its own format, so ES relies on the Common Information Model (CIM): an add-on with a set of data models that define common field names and tags for each domain, such as authentication, network traffic or endpoint activity. Normalising each new source to the CIM is part of onboarding it.

One detail matters for anyone comparing platforms. Splunk describes the CIM as a "search-time schema", which leaves the raw data intact and applies the normalisation when you search. It is flexible, but it also means that a source which is not mapped correctly is missing from the views built on it. Splunk's own documentation notes that dashboards and reports in CIM-aware apps such as ES display only the data normalised to the CIM tags and fields.

Correlation searches, now called detections

The heart of any SIEM is its rules. In ES these were long known as correlation searches. From ES 8.0 Splunk calls them detections: scheduled correlation searches or risk rules, written in SPL (Search Processing Language), that look for a pattern across security data, asset and identity lists, threat intelligence and third-party alerts.

When a detection matches, it can create a finding (the alert an analyst triages, stored in the notable index), adjust a risk score, or run a response action. Splunk ships a library of out-of-the-box detections grouped into analytic stories, and every detection can be edited and versioned. Analysts work the queue in Mission Control, which integrates the Incident Review page of earlier releases.

Risk-based alerting

Many signals are not worth an alert on their own, such as a login from an unusual country. Risk-based alerting (RBA) handles them as intermediate findings. They are written to a risk index against an entity, such as a user or a host, and never reach the analyst queue directly. A second detection adds them up over time and creates a finding only when the entity crosses a threshold. Done well, RBA is how Splunk teams cut alert volume. It also takes tuning: Splunk's documentation describes applying less risk to findings that turn out to be expected behaviour, instead of lowering the overall risk score.

How Splunk licensing works for SIEM

Splunk does not publish an ES price list: pricing is quoted per workload. What it does publish is how usage is measured, and that is what you need to model a budget.

  • ES adds no capacity. Splunk's licensing page for ES gives an example: with a 1 GB Daily Indexing Volume licence plus ES, you can ingest 1 GB per day in total, and ES runs on that same data.
  • Ingest pricing is based on gigabytes of data ingested per day, with unlimited users and searches.
  • Workload pricing is based on the compute needed to run search and analytics: measured in Splunk Virtual Compute (SVC) units on Splunk Cloud Platform and in vCPUs on Splunk Enterprise. Splunk positions it for data you store but search rarely.
  • Activity-based pricing, available on Splunk Cloud Platform, meters both ingest and search activity.

Two practical consequences. First, the bill moves with either volume or search load, so a noisy source or a heavy set of scheduled detections both show up in the cost. Splunk offers Ingest Processor and Edge Processor to filter and route data before it is indexed, and that work is part of running Splunk well. Second, on Splunk Enterprise retention is something you configure and pay for in storage: each index rolls data to frozen when it reaches either its age limit or its size limit, and the indexer deletes frozen data by default unless you archive it.

When Splunk is a sensible SIEM choice

Splunk is a mature platform, and for some organisations it remains the right fit:

  • You use Splunk beyond security. If operations, observability and security already share one Splunk estate, the data and the skills serve several teams.
  • You need self-managed or isolated deployments. Splunk Enterprise runs in a private cloud or on-premises, including air-gapped environments.
  • You have deep SPL skills and a mature detection library. Years of tuned detections and RBA content are an asset, and they are not quick to rebuild elsewhere.

Why organisations review their Splunk SIEM

  • The version lifecycle. Splunk supports each minor version of ES for 24 months from release. Support for ES 7.3 ended on 28 February 2026, after an extension to give customers time to plan the move to ES 8.x, and Splunk's support policy sets 30 October 2026 as the end of support date for ES 8.0. ES 8 renamed core concepts (detections, findings, intermediate findings) and moved the analyst workflow into Mission Control. Teams still on 7.x face a project with retraining either way, which makes it a natural point to compare options.
  • Data growth. Every new cloud account or business application adds sources, and under ingest or workload pricing that growth shows up in the renewal.
  • Operational load. On Splunk Enterprise, the indexers, search heads, storage and upgrades are your team's work. For a mid-market security team, platform maintenance competes with detection work.
  • Renewal timing. Term licences and annual subscriptions create a fixed window. A migration that starts too close to it often ends with a short-term renewal anyway.

Splunk and Google Security Operations: the differences that matter

When organisations look beyond Splunk, Google Security Operations (Google SecOps) is one of the platforms they compare. The differences worth testing are structural.

  • Cost model. Google states that "the pricing for all packages is based on ingestion volume" across its Standard, Enterprise and Enterprise Plus packages. Because the meter is ingestion, the number and breadth of searches your analysts run does not change the bill.
  • Retention. According to Google, "by default, Google retains 12 months of your data", extendable up to five years in the purchase order. The Standard package describes this as hot retention, searchable at no cost beyond ingestion, so there is no storage to size for that first year.
  • Normalisation. Google SecOps normalises telemetry into its Unified Data Model (UDM) as data is ingested, through parsers, rather than at search time. Detections and searches then work on the same consistent fields.
  • Detection language. Rules are written in YARA-L 2.0, which is section-based where SPL is a chain of piped commands. Google publishes a dedicated guide for teams moving from SPL, mapping commands such as stats and dedup to YARA-L sections.
  • Included content. Curated detection rules maintained by Google come as a subset in Standard and without limits from Enterprise upwards, which also adds UEBA and Gemini in Security Operations.
  • Data location. Google publishes the locations available for Google SecOps, including a Europe multi-region and single regions such as London, Frankfurt, Paris, Turin and Warsaw.

For a broader view of the criteria, including how Microsoft Sentinel compares, see our guide to choosing SIEM tools.

What to consider in a Splunk SIEM migration

A migration away from Splunk is less about moving data than about moving detection capability without a blind spot.

  1. Inventory what you actually use. List sources, sourcetypes, add-ons and daily volume per index. Then list the detections that produce findings someone works, which is usually a fraction of those enabled. That second list is the scope of the project.
  2. Map CIM to UDM. For each source, check that a Google SecOps parser exists and extracts the fields your detections rely on. A field that was normalised at search time in Splunk has to be present in UDM after ingestion.
  3. Rewrite detections, do not translate them line by line. SPL logic built on join, transaction or stats becomes event correlation on a shared field over a time window in YARA-L. Risk-based alerting content needs the most thought, because the scoring logic has to be redesigned rather than copied.
  4. Decide where history lives. Splunk retention ends when an index rolls to frozen, and frozen data is deleted unless archived. Before decommissioning, decide which historical data you keep, where, and for how long, based on your compliance and investigation needs.
  5. Run both platforms in parallel. Google SecOps includes a Splunk integration in its SOAR marketplace, with a connector that ingests notable events from Splunk ES. During coexistence, cases can flow into one queue while you validate the new detections against the old ones. Under Article 23(4) of Directive (EU) 2022/2555 (NIS2), entities in scope send an early warning within 24 hours of becoming aware of a significant incident, so a detection gap during cutover is a compliance issue as well as a security one.
  6. Plan around the licence term. Work back from the Splunk renewal date and leave room for the parallel run, not just for the technical work.
  7. Decide who runs it afterwards. Runbooks, rule ownership and out-of-hours cover need an owner on day one of the new platform, not after the first incident.

Where AmagisTech fits

AmagisTech is a Google Cloud partner and runs managed SOC services for mid-market organisations on Google Security Operations. We help teams assess their Splunk estate, plan the move, run the parallel phase and operate the platform afterwards. Our observability and SecOps service describes how we work, and our managed security services cover monitoring once the platform is live.

If you are approaching a Splunk renewal or an ES upgrade and want a second view on your options, talk to our team.

In short: Splunk works as a SIEM through Enterprise Security, a premium app on Splunk Enterprise or Splunk Cloud Platform. It normalises data with the CIM at search time, runs SPL detections that create findings, and uses risk-based alerting to cut noise. Licensing follows ingest, compute or activity. Organisations usually review it at an ES upgrade, a renewal or when data growth changes the cost, and a migration succeeds when detections, history and the parallel run are planned before the cutover.

Frequently asked questions

Splunk Enterprise and Splunk Cloud Platform index and search machine data. The SIEM is Splunk Enterprise Security, a premium app that runs on either platform and adds detections, findings, risk-based alerting, investigation and case management.
Splunk Enterprise is the self-managed platform that collects, indexes and searches data. Enterprise Security is the security app on top of it (or on Splunk Cloud Platform). You need the platform licence first, and Enterprise Security works on the data that licence already covers.
Splunk does not publish an Enterprise Security price list and quotes per workload. Usage is measured on ingest (GB per day), on workload (SVC units on Splunk Cloud Platform, vCPUs on Splunk Enterprise) or, on Splunk Cloud Platform, on activity-based pricing that meters both ingest and search.
Cisco. On 18 March 2024 Cisco announced that it had completed the acquisition of Splunk, for approximately USD 28 billion in equity value.
Yes. The work is mapping each source from the Splunk CIM to Google's Unified Data Model, rewriting the detections you actually use in YARA-L (Google publishes a guide for teams coming from SPL), deciding where historical data stays, and running both platforms in parallel before you switch Splunk off.

Let's talk about your security.

A 30-minute call to understand where you stand today and tell you, with no commitment, where it pays to start.

Book a call