#include <bits/stdc++.h>
using namespace std;
#define FOR(i,m,n) for(int i=(m);i<(n);++i)
#define REP(i,n) FOR(i,0,n)
#define ALL(v) (v).begin(),(v).end()
using ll = long long;
constexpr int INF = 0x3f3f3f3f;
constexpr long long LINF = 0x3f3f3f3f3f3f3f3fLL;
constexpr double EPS = 1e-8;
constexpr int MOD = 998244353;
// constexpr int MOD = 1000000007;
constexpr int DY4[]{1, 0, -1, 0}, DX4[]{0, -1, 0, 1};
constexpr int DY8[]{1, 1, 0, -1, -1, -1, 0, 1};
constexpr int DX8[]{0, -1, -1, -1, 0, 1, 1, 1};
template <typename T, typename U>
inline bool chmax(T& a, U b) { return a < b ? (a = b, true) : false; }
template <typename T, typename U>
inline bool chmin(T& a, U b) { return a > b ? (a = b, true) : false; }
struct IOSetup {
  IOSetup() {
    std::cin.tie(nullptr);
    std::ios_base::sync_with_stdio(false);
    std::cout << fixed << setprecision(20);
  }
} iosetup;

template <unsigned int M>
struct MInt {
  unsigned int v;

  constexpr MInt() : v(0) {}
  constexpr MInt(const long long x) : v(x >= 0 ? x % M : x % M + M) {}
  static constexpr MInt raw(const int x) {
    MInt x_;
    x_.v = x;
    return x_;
  }

  static constexpr int get_mod() { return M; }
  static constexpr void set_mod(const int divisor) {
    assert(std::cmp_equal(divisor, M));
  }

  static void init(const int x) {
    inv<true>(x);
    fact(x);
    fact_inv(x);
  }

  template <bool MEMOIZES = false>
  static MInt inv(const int n) {
    // assert(0 <= n && n < M && std::gcd(n, M) == 1);
    static std::vector<MInt> inverse{0, 1};
    const int prev = inverse.size();
    if (n < prev) return inverse[n];
    if constexpr (MEMOIZES) {
      // "n!" and "M" must be disjoint.
      inverse.resize(n + 1);
      for (int i = prev; i <= n; ++i) {
        inverse[i] = -inverse[M % i] * raw(M / i);
      }
      return inverse[n];
    }
    int u = 1, v = 0;
    for (unsigned int a = n, b = M; b;) {
      const unsigned int q = a / b;
      std::swap(a -= q * b, b);
      std::swap(u -= q * v, v);
    }
    return u;
  }

  static MInt fact(const int n) {
    static std::vector<MInt> factorial{1};
    if (const int prev = factorial.size(); n >= prev) {
      factorial.resize(n + 1);
      for (int i = prev; i <= n; ++i) {
        factorial[i] = factorial[i - 1] * i;
      }
    }
    return factorial[n];
  }

  static MInt fact_inv(const int n) {
    static std::vector<MInt> f_inv{1};
    if (const int prev = f_inv.size(); n >= prev) {
      f_inv.resize(n + 1);
      f_inv[n] = inv(fact(n).v);
      for (int i = n; i > prev; --i) {
        f_inv[i - 1] = f_inv[i] * i;
      }
    }
    return f_inv[n];
  }

  static MInt nCk(const int n, const int k) {
    if (n < 0 || n < k || k < 0) [[unlikely]] return MInt();
    return fact(n) * (n - k < k ? fact_inv(k) * fact_inv(n - k) :
                                  fact_inv(n - k) * fact_inv(k));
  }
  static MInt nPk(const int n, const int k) {
    return n < 0 || n < k || k < 0 ? MInt() : fact(n) * fact_inv(n - k);
  }
  static MInt nHk(const int n, const int k) {
    return n < 0 || k < 0 ? MInt() : (k == 0 ? 1 : nCk(n + k - 1, k));
  }

  static MInt large_nCk(long long n, const int k) {
    if (n < 0 || n < k || k < 0) [[unlikely]] return MInt();
    inv<true>(k);
    MInt res = 1;
    for (int i = 1; i <= k; ++i) {
      res *= inv(i) * n--;
    }
    return res;
  }

  constexpr MInt pow(long long exponent) const {
    MInt res = 1, tmp = *this;
    for (; exponent > 0; exponent >>= 1) {
      if (exponent & 1) res *= tmp;
      tmp *= tmp;
    }
    return res;
  }

  constexpr MInt& operator+=(const MInt& x) {
    if ((v += x.v) >= M) v -= M;
    return *this;
  }
  constexpr MInt& operator-=(const MInt& x) {
    if ((v += M - x.v) >= M) v -= M;
    return *this;
  }
  constexpr MInt& operator*=(const MInt& x) {
    v = (unsigned long long){v} * x.v % M;
    return *this;
  }
  MInt& operator/=(const MInt& x) { return *this *= inv(x.v); }

  constexpr auto operator<=>(const MInt& x) const = default;

  constexpr MInt& operator++() {
    if (++v == M) [[unlikely]] v = 0;
    return *this;
  }
  constexpr MInt operator++(int) {
    const MInt res = *this;
    ++*this;
    return res;
  }
  constexpr MInt& operator--() {
    v = (v == 0 ? M - 1 : v - 1);
    return *this;
  }
  constexpr MInt operator--(int) {
    const MInt res = *this;
    --*this;
    return res;
  }

  constexpr MInt operator+() const { return *this; }
  constexpr MInt operator-() const { return raw(v ? M - v : 0); }

  constexpr MInt operator+(const MInt& x) const { return MInt(*this) += x; }
  constexpr MInt operator-(const MInt& x) const { return MInt(*this) -= x; }
  constexpr MInt operator*(const MInt& x) const { return MInt(*this) *= x; }
  MInt operator/(const MInt& x) const { return MInt(*this) /= x; }

  friend std::ostream& operator<<(std::ostream& os, const MInt& x) {
    return os << x.v;
  }
  friend std::istream& operator>>(std::istream& is, MInt& x) {
    long long v;
    is >> v;
    x = MInt(v);
    return is;
  }
};
using ModInt = MInt<MOD>;

template <typename Abelian>
struct FenwickTree {
  explicit FenwickTree(const int n, const Abelian ID = 0)
      : n(n), ID(ID), data(n, ID) {}

  void add(int idx, const Abelian val) {
    for (; idx < n; idx |= idx + 1) {
      data[idx] += val;
    }
  }

  Abelian sum(int idx) const {
    Abelian res = ID;
    for (--idx; idx >= 0; idx = (idx & (idx + 1)) - 1) {
      res += data[idx];
    }
    return res;
  }

  Abelian sum(const int left, const int right) const {
    return left < right ? sum(right) - sum(left) : ID;
  }

  Abelian operator[](const int idx) const { return sum(idx, idx + 1); }

  int lower_bound(Abelian val) const {
    if (val <= ID) [[unlikely]] return 0;
    int res = 0;
    for (int mask = std::bit_ceil(static_cast<unsigned int>(n + 1)) >> 1;
         mask > 0; mask >>= 1) {
      const int idx = res + mask - 1;
      if (idx < n && data[idx] < val) {
        val -= data[idx];
        res += mask;
      }
    }
    return res;
  }

 private:
  const int n;
  const Abelian ID;
  std::vector<Abelian> data;
};

int main() {
  int n; cin >> n;
  vector<int> b(n), c(n); REP(i, n) cin >> b[i] >> c[i];
  vector<int> ord(n);
  iota(ALL(ord), 0);
  ranges::sort(ord, {}, [&](const int i) -> int { return c[i]; });
  vector<int> a(n + 1, 0);
  ranges::copy(b, next(a.begin()));
  ranges::sort(a);
  a.erase(unique(a.begin(), a.end()), a.end());
  const int m = a.size();
  FenwickTree<int> num(m);
  FenwickTree<ModInt> r_sum(m), len_sum(m), mus_r(m), mus_nel(m);
  ModInt ans = 0;
  for (const int i : ord) {
    const int index = distance(a.begin(), ranges::lower_bound(a, b[i]));
    const int always = num.sum(index, m), len = c[i] - b[i] + 1;
    ans += ModInt::raw(always) / len;
    ans += r_sum.sum(index) / len - len_sum.sum(index) * (b[i] - 1) / len;
    ans -= mus_r.sum(index) / len - mus_nel.sum(index) * (b[i] - 1) / len;
    num.add(index, 1);
    r_sum.add(index, ModInt::raw(c[i]) / len);
    len_sum.add(index, ModInt::inv(len));
    const int right = distance(a.begin(), ranges::lower_bound(a, c[i])) - 1;
    mus_r.add(right, ModInt::raw(c[i]) / len);
    mus_nel.add(right, ModInt::inv(len));
  }
  cout << (-ans + n * (n - 1LL) / 2) * ModInt::fact(n) / 2 << '\n';
  return 0;
}