The Intermediary – March 2026 - Flipbook - Page 79
T E C H N O L O GY
Opinion
No technology
should be an island
F
or many lenders, the
discussion around cloud
migration still begins
with infrastructure. It
starts with data centres,
hosting locations and
the practical question of how to move
existing systems into a more modern
environment without introducing
unnecessary risk.
Given the longstanding emphasis
on resilience, auditability and
operational control within
mortgage lending, that instinct is
entirely understandable.
Where the conversation can become
constrained, however, is when cloud is
framed primarily as a change of venue
rather than a change of behaviour.
‘Li and shi’ approaches that
relocate legacy platforms into public
cloud environments with minimal
architectural change are oen
presented as prudent first steps. They
appear to offer modern hosting while
preserving familiar operating models.
In some contexts, that may be a
rational transitional choice. But it is
important to be clear about what it
does and does not achieve.
Cloud native architecture is
not simply an implementation
preference. It represents a different
operating philosophy.
The substantive advantage of cloud
lies less in where systems reside and
more in how they are designed to
respond under real world conditions.
A genuinely cloud native platform
assumes volatility as normal.
Volume surges, product
launches, distribution shis and
seasonal peaks are not treated as
exceptional stress events requiring
structural reconfiguration. They
are accommodated through
elastic capabilities built into the
architecture itself.
That said, elasticity does not remove
the need for disciplined planning.
Lenders still need clear visibility
of expected origination volumes,
refinancing waves and servicing
activity. Cloud does not negate those
responsibilities. It complements them.
Hosting a mortgage platform on
Amazon Web Services, in isolation,
does not create that outcome.
Designing the platform to exploit
elasticity as a first principle does.
Capacity can expand within agreed
parameters when demand increases
and contract when it recedes.
Infrastructure is provisioned
programmatically, but within
defined performance and commercial
guardrails. Performance tuning and
failover are embedded behaviours
rather than emergency responses.
This shi has economic as well as
operational implications. Traditional
environments require capacity to be
provisioned for projected peaks, oen
well in advance. That model can lock
in cost and constrain flexibility. An
elastic model allows infrastructure to
flex in line with actual usage.
Architectural mechanism
However, this flexibility works
most effectively when paired with
transparent volume expectations and
collaborative planning. Auto scaling
is not a substitute for forecasting. It
is an architectural mechanism that
makes forecasting less brile and less
operationally disruptive.
Li and shi platforms rarely
deliver this outcome because they
carry forward assumptions from a
different era. Even when hosted in the
cloud, these systems oen rely heavily
on operators to predict demand and
coordinate scaling events.
The physical location changes
but the operating mindset does not.
A cloud native model, by contrast,
embeds scaling logic into the platform
while still operating within clearly
defined commercial and performance
boundaries agreed with the lender.
There is a further distinction that
becomes increasingly relevant in
regulated markets such as mortgages.
JERRY MULLE
is UK managing director
at Ohpen
Cloud native platforms are designed
to evolve continuously. A true cloud
native partner actively reviews and
incorporates relevant improvements
into the service stack.
Lied legacy systems tend to remain
comparatively static. Because they are
tightly coupled and not designed for
modular change, even incremental
adjustments can introduce
disproportionate risk. Over time, this
can recreate the rigidity that cloud
migration was intended to address.
For lenders operating under
regulatory expectations, the
implications are significant. Service
availability, auditability and data
integrity are not discretionary.
Mortgage lending demands
caution and operational continuity is
paramount. But it does suggest that
the cloud conversation benefits from a
broader framing.
If cloud is viewed primarily as an
infrastructure exercise, the outcome
will likely be incremental. If it is
viewed as an opportunity to align
system behaviour with the volatility
and pace of modern lending markets
the implications are more strategic.
In that context, li and shi is
best understood as relocation. It
may address certain infrastructure
constraints, but it does not remove
embedded assumptions about capacity,
failure or change. Cloud native
architecture, by contrast, is designed
to absorb volatility and evolve
continuously alongside the business.
The cloud is not a destination to
be reached. It is a capability to be
exercised. Realising its full value
depends on designing systems and
governance frameworks that are built
to exploit it fully. ●
March 2026 | The Intermediary
77
